{"_id":"pi-llama-cpp","_rev":"32-4a829dcc83c1aa8c21d70c5aaca2c93e","name":"pi-llama-cpp","dist-tags":{"latest":"0.16.0"},"versions":{"0.1.0":{"name":"pi-llama-cpp","version":"0.1.0","keywords":["pi-package","pi-extension","llama-cpp","llama.cpp"],"_id":"pi-llama-cpp@0.1.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"pi":{"extensions":["./index"]},"dist":{"shasum":"935ed3702bc4d00b669c883ddc69712c40927903","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.1.0.tgz","fileCount":18,"integrity":"sha512-UbR14YXj1yktzM9Z2lt0xjJR2rteBy7wQ9HcRVgoUBvIl3n1jXB5ZZs+xKH/CfTtlsokFG+IJue284wtynw8Ng==","signatures":[{"sig":"MEYCIQDlpESRdmIKIf7epA2OeuoMvT47cuMVwIWHw2lfr7cRCQIhAMljABUEsRxmmaW7Ts+H5BfOR2HutqTwcZR92ZbH1Hj5","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":23168},"gitHead":"57746297afaeaa20f9b1808adecb3ec1377646c9","_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"_npmVersion":"11.13.0","description":"Pi extension for llama.cpp integration. Supports both router and single modes","directories":{},"_nodeVersion":"20.20.2","_hasShrinkwrap":false,"peerDependencies":{"@mariozechner/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.1.0_1777250219125_0.7302767656908304","host":"s3://npm-registry-packages-npm-production"}},"0.1.1":{"name":"pi-llama-cpp","version":"0.1.1","keywords":["pi","pi-package","pi-extension","llama-cpp"],"_id":"pi-llama-cpp@0.1.1","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"pi":{"extensions":["./index.ts"]},"dist":{"shasum":"0bd35d261a5f3cdfbc6734f9511b780578013c14","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.1.1.tgz","fileCount":18,"integrity":"sha512-yTEBSsbYIAxeuyFWMdjj0RHE2QnT4Oo84B98UIRh/ld0SwwZ7BYa5okLGE2BQ6Oek6gwS9nM8h7XhdFTShluzg==","signatures":[{"sig":"MEYCIQCTQ/CweQcOpHUD2583gJSF0kNTtDkCa3Tdo89bDE2xRgIhALbdPNibqKiOw+gPhgPPpZxfHX97ZrbvPCGTk466vSJ7","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":23164},"gitHead":"60d7e6011cf7f087bd06a8afe99ed93bb99b41f3","_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"_npmVersion":"11.13.0","description":"Pi extension for llama.cpp integration. Supports both router and single modes","directories":{},"_nodeVersion":"20.20.2","_hasShrinkwrap":false,"peerDependencies":{"@mariozechner/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.1.1_1777250520944_0.6731173217633333","host":"s3://npm-registry-packages-npm-production"}},"0.1.2":{"name":"pi-llama-cpp","version":"0.1.2","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.1.2","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./index.ts"]},"dist":{"shasum":"663552390161589a8f83095b4d15ba6d02c1425e","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.1.2.tgz","fileCount":18,"integrity":"sha512-Ht7OR51qsX1+tb9/Neeq0WnGdpAg10eTtY7qw3ys0l4JsRNIS1y8puPPnHGX/9QdG0KYoh6cTJpb3TKfURjJsg==","signatures":[{"sig":"MEUCIQCV5U+ry2iFipMcd6NJNWyEsfLDSpd1tBE8JLm1WAJUUAIgdYVSHKqf71hnJEoVX/IcBRy+t0z/Cj3Z+42Mx6WN1xM=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":23429},"gitHead":"a3c2e43cfed972cfa27a1a2c6fc9f3f1c1e396dd","_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.13.0","description":"Pi extension for llama.cpp integration. Supports both router and single modes","directories":{},"_nodeVersion":"20.20.2","_hasShrinkwrap":false,"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.1.2_1777253456393_0.19033580432814623","host":"s3://npm-registry-packages-npm-production"}},"0.2.0":{"name":"pi-llama-cpp","version":"0.2.0","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.2.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"dfc0cf1dd1418c083972c88891c6de3be73d813c","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.2.0.tgz","fileCount":20,"integrity":"sha512-Dgn13HYHJ2X5MOw4XSG6oSvgjp++q9uWcrkfVu4frfcbnU09ZNumrx2WK44x6uR6ZBE0mH2bKk7jdqfPvoGGCA==","signatures":[{"sig":"MEUCIQDn3BMB7AudyOEIm4RWC8j3fTjjtPT4K1wwC9XGIEwa3AIgQfRFbi453IGB96fHAdPClP+DlP9tTycvRRV3dcybKx4=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":26791},"gitHead":"29e5050689dbbc009f523f7f4a7dc743948ab55e","_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.13.0","description":"Pi extension for llama.cpp integration. Supports both router and single modes","directories":{},"_nodeVersion":"20.20.2","_hasShrinkwrap":false,"devDependencies":{"@types/node":"^25.6.0","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@mariozechner/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.2.0_1777600825594_0.024798229977759823","host":"s3://npm-registry-packages-npm-production"}},"0.2.1":{"name":"pi-llama-cpp","version":"0.2.1","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.2.1","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"82361ff193e2f2386126792d0ad06476eaa7693e","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.2.1.tgz","fileCount":20,"integrity":"sha512-Xthhpdq/XbncXZfkOEzMPZ0dzoDEMnh2X+5aVJsPnyiqvaccaB0MdOKJjvkRLLK/IQRsM78dZT5s86kcVTzxiw==","signatures":[{"sig":"MEYCIQCTTcdh2zeDVIuGTrtwxzthcBCtbdw0vEytZA9Ee3y8ogIhAKpEIlXxL6JgoH7uWfeC+uO4ONnktQRrVYPSvhJ7YEhu","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":28436},"gitHead":"ca335b8572b2d31dc847ade2618666d5d72adfc0","_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.13.0","description":"Pi extension for llama.cpp integration. Supports both router and single modes","directories":{},"_nodeVersion":"20.20.2","_hasShrinkwrap":false,"devDependencies":{"@types/node":"^25.6.0","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@mariozechner/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.2.1_1777606717407_0.7414530252076541","host":"s3://npm-registry-packages-npm-production"}},"0.2.2":{"name":"pi-llama-cpp","version":"0.2.2","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.2.2","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"439aa94e04fdc4874dede9c91320e533a7172f75","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.2.2.tgz","fileCount":22,"integrity":"sha512-kf0GM8Qbkdm+32FBe29Q0LjRt189v1wK/5qxUpBAbC3bwVQa6HJ+T/cOHR6yWoVIYhdvhHA0OIEBVtH9uidI4A==","signatures":[{"sig":"MEUCIHN7/IXyFHod4pKBsDdOg9fKlXmMgVLTp/UKR7y3qjngAiEAmjmzAwptqB6MVyOpud5Toc45Eou0+/4BPzDOwcC88MY=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":31753},"gitHead":"5bb392ea1bbf5ea0c152a467c850b02e7d17efdf","_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.13.0","description":"Pi extension for llama.cpp integration. Supports both router and single modes","directories":{},"_nodeVersion":"20.20.2","_hasShrinkwrap":false,"devDependencies":{"@types/node":"^25.6.0","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@mariozechner/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.2.2_1777669610861_0.550274823492853","host":"s3://npm-registry-packages-npm-production"}},"0.2.3":{"name":"pi-llama-cpp","version":"0.2.3","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.2.3","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"e96c720f462c69bc7b7a112329295bfdda6677c0","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.2.3.tgz","fileCount":27,"integrity":"sha512-nbua4Y9+/FGeJbrbwnSBZTi3Jhe9xTJIjyl2OGjh8VOkwBqL86Q52rNwn6KUQueyEPqA08w4YjwPwM0cpyXNzw==","signatures":[{"sig":"MEUCIQC+BFGZpldZvqTsNvmdEVd12SL5mDSU3+Fj6S3FPnRS3AIgUovmIU2cqHD+BKevBQsiexEBuA0YU/HB8kEDET8F3Fc=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":49422},"gitHead":"a1001ccfe942b767ab2bd77dd058325bc0da90d7","scripts":{"test":"vitest","test:run":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.13.0","description":"Pi extension for llama.cpp integration. Supports both router and single modes.","directories":{},"_nodeVersion":"20.20.2","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.5","@types/node":"^25.6.0","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@mariozechner/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.2.3_1777693424413_0.9355321276889836","host":"s3://npm-registry-packages-npm-production"}},"0.3.0":{"name":"pi-llama-cpp","version":"0.3.0","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.3.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"325d8d6230ac949583b28c806fe176e359a03955","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.3.0.tgz","fileCount":27,"integrity":"sha512-fbbFVUzgkqxuU5ICayMmcLWDVCt/mY5EiY+Sxglo6JKPD+LBPTmyp4+XVXfhj7m+2LxWhWPsFVd4USIjiRhkxA==","signatures":[{"sig":"MEQCIFzLlGJRbBube6Dy8hwbRgdDcVjl0gj2EdGkwJfVj47kAiA7d0CiZ7wj2M5ZqGXFWx+eSi8AqGH9hyUnKrBoIQDlgQ==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":54086},"gitHead":"ba6887589ac927ea3424eb5b65bb0f8353b9e400","scripts":{"test":"vitest","test:run":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.14.1","description":"Pi extension for llama.cpp integration. Supports both router and single modes.","directories":{},"_nodeVersion":"20.20.2","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.5","@types/node":"^25.6.0","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.3.0_1778351245529_0.23009167218595894","host":"s3://npm-registry-packages-npm-production"}},"0.3.1":{"name":"pi-llama-cpp","version":"0.3.1","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.3.1","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"194efcce1265170a224ff629abc927735926fa9e","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.3.1.tgz","fileCount":27,"integrity":"sha512-1F5t4QX2mJ4vVF6+L5J5JsxewGWFHQgQhcQylIWoD9Awvg2eD2pIsAFHATe4gXaz8eeShtsaaf3RCnaZc+cYOA==","signatures":[{"sig":"MEUCIQCy80FXx6ukRMi45+OdmgHO1P28ax0JKhXnKjT6fHwQPAIgCKOX3/7syYcdwMeOHJhDipoY/7/jxl0fc1RjQV3g0rk=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":54574},"gitHead":"1a050bde41c263d8b546d81c4f1cb8bad350615f","scripts":{"test":"vitest","test:run":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.14.1","description":"Pi extension for llama.cpp integration. Supports both router and single modes.","directories":{},"_nodeVersion":"20.20.2","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.5","@types/node":"^25.6.0","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.3.1_1778356993234_0.5041026011405374","host":"s3://npm-registry-packages-npm-production"}},"0.3.2":{"name":"pi-llama-cpp","version":"0.3.2","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.3.2","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"457df1b6051aa36e7eb4d3a66ea701191d0a237d","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.3.2.tgz","fileCount":27,"integrity":"sha512-8uxDiococmu7Cpixf2TODqD3gbuDAevmIfqOZ9EkZM8Y1WvK1mEmgAwuYTmlId2eXlUk64phGCX9HVt+f1qDNQ==","signatures":[{"sig":"MEUCIBSI+J9o3Wd3aiaJW0hK5UduqQ4YWvQG3zW78NyWx0NPAiEA1X2PK4OZPaSTB4/X4GwQNXqmaRjRsqXDTVVQHS3DEdo=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":55947},"gitHead":"cd0c98aa708d0d44d81d0cac4364464a5ea3561b","scripts":{"test":"vitest","test:run":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.14.1","description":"Pi extension for llama.cpp integration. Supports both router and single modes.","directories":{},"_nodeVersion":"20.20.2","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.5","@types/node":"^25.6.0","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.3.2_1778386070706_0.8037181943809599","host":"s3://npm-registry-packages-npm-production"}},"0.3.3":{"name":"pi-llama-cpp","version":"0.3.3","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.3.3","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"2fe3cb65e46b1444222c30fc4c0a895363bdb871","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.3.3.tgz","fileCount":27,"integrity":"sha512-7scCZ09ucxdR888UD5u/mYRaCE1U8x8lcngbfIEiKlCI7/4S9r6Crl00IeDuQ/LCTNPSHcfHgNVZjbuswy5ccA==","signatures":[{"sig":"MEQCICOOVWQUiBv/XyhZnHVG6+IuHsVzOWqDETNzVrYJ/kRNAiAJ7T8xx4gBCQnZlU1iqivX8xVwTXUIx9tynGoyC4c2ww==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":56744},"gitHead":"23e513184e629ba35fcc4c36c51af9c340c7a6e9","scripts":{"test":"vitest","test:run":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.14.1","description":"Pi extension for llama.cpp integration. Supports both router and single modes.","directories":{},"_nodeVersion":"20.20.2","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.5","@types/node":"^25.6.0","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.3.3_1778397503551_0.12734776506510137","host":"s3://npm-registry-packages-npm-production"}},"0.3.4":{"name":"pi-llama-cpp","version":"0.3.4","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.3.4","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"7c5de725d04a1d28249fafa3bd8c5c66b86e80b4","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.3.4.tgz","fileCount":27,"integrity":"sha512-IIRVrzKjy0H6y3QrJEqCaVL4id4S1oKVkPK16Os5OJ6I9gsvzI3u87Ue8Ei/bFS8oXqUyB5K3wcVRf0sC+oVpg==","signatures":[{"sig":"MEYCIQCj7/xgKaz77vydXfA3YjjUn4RJkVI7uftovnQAownoOQIhAM/2CU/l5/Jvd9x05n7aUyh//M1EgWh6kQ5JOL/Hxh5h","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":56031},"gitHead":"ba306b9c56764ae36e488d4c4bdf46be75c6a4a9","scripts":{"test":"vitest","test:run":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.14.1","description":"Pi extension for llama.cpp integration. Supports both router and single modes.","directories":{},"_nodeVersion":"20.20.2","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.5","@types/node":"^25.6.0","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.3.4_1778398767228_0.13418806524115467","host":"s3://npm-registry-packages-npm-production"}},"0.4.0":{"name":"pi-llama-cpp","version":"0.4.0","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.4.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"0f12578fdb03d7fb639660e4c2aa5fd1e0f5a61b","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.4.0.tgz","fileCount":29,"integrity":"sha512-xdlAUZSdBgcomaWxsp6MVvbB+ugB9BMcAzTpzmJKyGeZllTe+l5dnS9uGeJpoutn3/61oY5WjGjNingJVdd0BA==","signatures":[{"sig":"MEUCIH34pYBL8jcUh4UDU3WXn3sN+R/cFzA6yVxYMASyMy/pAiEAhhz91NLDkIdVly8QzSBCE4VhAwN8M0p+MC4Oo2n9bQo=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":63548},"gitHead":"fc7a8f1d4628424ea46797b09ab4f78a591d10bc","scripts":{"test":"vitest","test:run":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.14.1","description":"Pi extension for llama.cpp integration. Supports both router and single modes.","directories":{},"_nodeVersion":"20.20.2","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.5","@types/node":"^25.6.0","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.4.0_1778911138077_0.31871735928176737","host":"s3://npm-registry-packages-npm-production"}},"0.5.0":{"name":"pi-llama-cpp","version":"0.5.0","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.5.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"20bd934b66f65e5c65c892157360d670d7f90326","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.5.0.tgz","fileCount":29,"integrity":"sha512-8k+cjVzHs4JSxgR+7GDcnLMysim9mW9/UfkJ0tGPlRBgETwZnQvxAcZTIRnr60xmwfW038lYnkWVa+6/dC9SDw==","signatures":[{"sig":"MEYCIQCMcFAKkCEfIhUddmuW9aQpNLehWftU77sY8/k7X/ZY2wIhAK52dXgPZOcuEEY3KyI3OUJG/vhQeu8nOAZgBuMNnU/1","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":70993},"gitHead":"8e6691fe3ef80d3bef12c34dc2fd4b341bb9c891","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.14.1","description":"Pi extension for llama.cpp integration. Supports both router and single modes.","directories":{},"_nodeVersion":"24.14.1","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.7","@types/node":"^25.9.1","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.5.0_1779510646642_0.9789339878300527","host":"s3://npm-registry-packages-npm-production"}},"0.5.1":{"name":"pi-llama-cpp","version":"0.5.1","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.5.1","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"9f39cf08d29b8bb0e498912c69ac3f5696d7ec98","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.5.1.tgz","fileCount":29,"integrity":"sha512-fO7UBrpF2XYCrgVdHiTCNyy3QlDSEX+P/a19QLw0UvSuWoJHPo60kFpDVJg0lKiUgIYdRyGooMZBWS4HPp0yiw==","signatures":[{"sig":"MEQCIHsIBr41tEHe/yuPQvP1E9yXfsFUwR11rvS05cSTqeULAiADurSKwaZJSKHo2ARsDxJoCiqFgBYCijHxeO9ppu6I3Q==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":71139},"gitHead":"7b2f2785ba8bf5e41e4736dac1d679c93ab2b051","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.14.1","description":"Pi extension for llama.cpp integration. Supports both router and single modes.","directories":{},"_nodeVersion":"24.14.1","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.7","@types/node":"^25.9.1","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.5.1_1780254217114_0.9154038708622787","host":"s3://npm-registry-packages-npm-production"}},"0.6.0":{"name":"pi-llama-cpp","version":"0.6.0","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.6.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"fdee5127871aa64fc10308d29ec5ee24af5e118d","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.6.0.tgz","fileCount":32,"integrity":"sha512-3+LiuGRQBAV4q3TBVn03FvWbWig/fZ6sTKvH8GAiAWh5Th9D2d9awU9vnoHV/8fOSfggVY42jQKANKjRd1O0KA==","signatures":[{"sig":"MEUCIQCqQJfxSob9W+h0lgp2pid82QX2tQq/JTc3MYTivVCSmwIgZYIPxbyO4m2i2rx2WMcBHmW/YgpLRvG+wZySYy1Hp2c=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":85092},"gitHead":"7f15255a47ab4a3a5f4c533edcd2a6e766d020c0","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.16.0","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.14.1","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.8","@types/node":"^25.9.1","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-tui":"*","@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.6.0_1780768428263_0.060813766941681724","host":"s3://npm-registry-packages-npm-production"}},"0.7.0":{"name":"pi-llama-cpp","version":"0.7.0","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.7.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"771315540da17d10b6c63f3dc6d4cd76263488f2","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.7.0.tgz","fileCount":35,"integrity":"sha512-AsQ7zWFLwKUjrEOTEG0ICXeEGGi3eDmB7ckKXSP3qXj2Nspyj5PiU2sAH2sTYlZWJwpcJ/+IasL/VWgyKwwYZA==","signatures":[{"sig":"MEUCIANLxGgQO3XGFekGFvaMNi0Nele42OVBJO1Mh2WNBldRAiEAlGN53+19KGvppvkWind3WXZdYaKe5dDpZzdMF3nX1D4=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":101857},"gitHead":"d2c95778fb38ef08347071f2e762619ce80e3462","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.16.0","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.16.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.8","@types/node":"^25.9.3","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-tui":"*","@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.7.0_1781331301550_0.29246052694501357","host":"s3://npm-registry-packages-npm-production"}},"0.7.1":{"name":"pi-llama-cpp","version":"0.7.1","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.7.1","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"a57ffab173cd608af5d600a369aac81427cbbfa3","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.7.1.tgz","fileCount":35,"integrity":"sha512-3Q9vXEPjACWhQLwChXuCxQj4vtkfsMQC5lDdFsGDoYn7fB0zzvC0HfkRedorWcqDA3/Ps6QQSjTczmqSPOcT5g==","signatures":[{"sig":"MEYCIQCIL34enYpNh6gvz4m2LsxTc6fCQG0g+F4OOVqou2cWFgIhAJNIPkNgVMFD7LpSY4nOUn7918QneebYsUWS15CwbXbT","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":101838},"gitHead":"83381cc7bb71a82e5319329f73ed39ca2d1a18dd","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.16.0","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.16.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.8","@types/node":"^25.9.3","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-tui":"*","@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.7.1_1781332371550_0.729470318454011","host":"s3://npm-registry-packages-npm-production"}},"0.7.2":{"name":"pi-llama-cpp","version":"0.7.2","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.7.2","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"07eae72c7973fcf910944d335d8891c2db345b1a","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.7.2.tgz","fileCount":37,"integrity":"sha512-VhfNRCfkk4+J5SQTTxmV+bomASlOmtcdsAEb/afPv8+uraG3MPU62bvyJVcfYDeU3oH/H8Xe6nnNFg/VEp2GHw==","signatures":[{"sig":"MEUCIDqbgKjiY85gkqqMiaJRbv2zxAcpDbeD7tpGGG58v6S5AiEApp/UmhVn2kMIFX7Z3kU4kSBOB99Oj1TI741/392ak4E=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":104889},"gitHead":"def22c3a8ebcf90f7e9442efb089a22fae8058cd","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.16.0","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.16.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.9","@types/node":"^26.0.0","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-tui":"*","@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.7.2_1781930618180_0.6548478921555239","host":"s3://npm-registry-packages-npm-production"}},"0.8.0":{"name":"pi-llama-cpp","version":"0.8.0","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.8.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"f143edea5852b62b545226fe57e0ee02a8a699dc","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.8.0.tgz","fileCount":42,"integrity":"sha512-0j2lxlyV62ukKExgH058oObvSPbUdfP2ylrPp+jyCqYZFTufnKMO3Gz/zYB9w2zdXsma/+YNefJKE8DjMDQwgQ==","signatures":[{"sig":"MEUCIQDYPehAsFqYba/UP9Aau9Sm+R6gUg3LffyhUAXolV6NPQIgTUJVZf5kd5X3l/cix5qbp62t8cOBOfXFdjxxLFQPn4U=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":133400},"gitHead":"2e6a11d6729d3312e9ae2c39ab2e474ea9a0efe4","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.16.0","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.16.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.9","@types/node":"^26.0.1","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-tui":"*","@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.8.0_1782620501187_0.19605858867867676","host":"s3://npm-registry-packages-npm-production"}},"0.8.1":{"name":"pi-llama-cpp","version":"0.8.1","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.8.1","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"1aa4958b12a2e0fce6d63c8f8032d36ef6bd68c7","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.8.1.tgz","fileCount":41,"integrity":"sha512-zJRizp8GFu8AHeAFvo7oFtm5NOA+4ZYwecgJTb3vEgxrktc+XoXdU9OKtF1n24sirYmUUazgv4uPnooHGDlUfg==","signatures":[{"sig":"MEUCIGYE3Q7EmNLMrxNf5jXGk0SK73gM1fKTqiQx0l/E9Q/mAiEAy+JX+rw8WULZYMwXw3VqVqTMUSyTcK0gBJ0AXXoAZfU=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":121263},"gitHead":"bfae25150ebc9d4682a7620cc8c9e2b477a3c7f2","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"11.16.0","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.16.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.9","@types/node":"^26.0.1","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-tui":"*","@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.8.1_1782622274431_0.3631607593669848","host":"s3://npm-registry-packages-npm-production"}},"0.8.2":{"name":"pi-llama-cpp","version":"0.8.2","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.8.2","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"a24c2f7d6e0c36f11c2f0ee197c57c889182f950","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.8.2.tgz","fileCount":41,"integrity":"sha512-Sp+mMheT4ySXv1Eg42exzjrNCC27tk4ax2bILIFXUihr9AvpUcJLMRdUfm27lrRfgRUm4DeqW6tM1iY9k/OxFQ==","signatures":[{"sig":"MEUCIG2vnQawXwDCobc7aup9YDeK8gHNIKVoi5XYkP3lkfx1AiEApxVqTvmsmUtYRQfQ7IIBkepetVl89jjPGmQLR8o/eZA=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":121014},"gitHead":"efe8305fc78f5487a90bea3144e94561128f91ef","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"12.0.1","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.16.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.10","@types/node":"^26.1.1","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-tui":"*","@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.8.2_1784225807475_0.03379164025368642","host":"s3://npm-registry-packages-npm-production"}},"0.9.0":{"name":"pi-llama-cpp","version":"0.9.0","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.9.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"4cb5bdc755d5d82b5fbd312956dae9abc3de1fad","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.9.0.tgz","fileCount":40,"integrity":"sha512-YbBK4RWTdncZL6SEuGPdKmBALyfGCkHdXu251rNGOjfbfaIbWPVA8/E9+Luc4OPSdCRcXeIX3aN+bRBpeRJS2A==","signatures":[{"sig":"MEYCIQCEwrbanhru8OtiNMHNcAP6Z50+2+u5cjYPQYJILRFZtgIhAO5PUrVKq+Ev19VCi+DJZ/2rSh7RTR/HrVqy6pTe61Nm","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":121236},"gitHead":"164a17c747e91230ac5fc532f0a628f86dcfa258","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"12.0.1","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.16.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.10","@types/node":"^26.1.1","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-ai":"*","@earendil-works/pi-tui":"*","@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.9.0_1784237014087_0.10808055209492484","host":"s3://npm-registry-packages-npm-production"}},"0.9.1":{"name":"pi-llama-cpp","version":"0.9.1","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.9.1","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"9c96eb7392fe424dee9fc9f2faaf2b1b8fe36b20","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.9.1.tgz","fileCount":40,"integrity":"sha512-iDOuYsS2EyGwGSGdv0gW37zg+xnChvZm6Msd5NAQg+eP/kJXv7sK0AEV6suDgIrUbMmP+uSxSUcWZeI6V9/flA==","signatures":[{"sig":"MEQCIHKe9WfphFbDF6AuDpqGvsqEKRXhcrV5Xnm94WqaXBufAiAoejPFHcCQ5sBMNDdeouZ5UkTtkEq9UBfRnUgvAGxmVg==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":121517},"gitHead":"ad26b84d3c0adff81494330de470a8b762dd6364","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"12.0.1","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.16.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.10","@types/node":"^26.1.1","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-ai":"*","@earendil-works/pi-tui":"*","@earendil-works/pi-coding-agent":"*"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.9.1_1784594886047_0.7665439402473506","host":"s3://npm-registry-packages-npm-production"}},"0.9.2":{"name":"pi-llama-cpp","version":"0.9.2","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.9.2","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"2a464eeebd93e34646ef5a3a202b405054b585a1","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.9.2.tgz","fileCount":40,"integrity":"sha512-B2+oitxi/vc1GIvJLsR+wGjA2hnTB1YX0lUHs4TrfYN0fel3b3X6JdrmeFjv/VD6ZMldrXEFau9cdBdrhFkj/A==","signatures":[{"sig":"MEUCIG5CCPtQNyjXgA4WP0QZHo4LwoDNTVP4+Yn6oTpOT0LCAiEA0FCXvZe27PznJKru6v/ueiEHoIxkiwPDh2n777SKqQA=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":122410},"gitHead":"74272c2b2daa944376f818e1cc50da455c57e5fe","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"12.0.2","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.18.1","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.11","@types/node":"^26.2.0","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-ai":">=0.84.0","@earendil-works/pi-tui":">=0.84.0","@earendil-works/pi-coding-agent":">=0.84.0"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.9.2_1787264187474_0.5995607010384658","host":"s3://npm-registry-packages-npm-production"}},"0.10.0":{"name":"pi-llama-cpp","version":"0.10.0","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.10.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"a52ba82090439812ec279cd97cfd60673d6472f6","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.10.0.tgz","fileCount":42,"integrity":"sha512-Zsi6T1c18yWQoa1akfAVazT3bSxxTg5P8fRv22hSLiJsNEZGFzx1b8XypGkx57JFvuCDzrkzDou+JOCkZb3y+w==","signatures":[{"sig":"MEUCIQCH8Dbb2LCoKvELlxbjDXA8PfRO1rS8vpNeO1wg449E3QIgBI2PXAjF0jME5e16ylwpjmtyOQhXpEzNp+dupKi/ZvU=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":149441},"type":"module","gitHead":"7f9f5ff028149f18b8815a519535ea86afa981d2","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"12.0.2","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.19.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^4.1.11","@types/node":"^26.4.0","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-ai":">=0.84.0","@earendil-works/pi-tui":">=0.84.0","@earendil-works/pi-coding-agent":">=0.84.0"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.10.0_1788041821884_0.700808944617036","host":"s3://npm-registry-packages-npm-production"}},"0.11.0":{"name":"pi-llama-cpp","version":"0.11.0","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.11.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"afed0fda08d05ecb5954c42fae46e44586ca5fb0","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.11.0.tgz","fileCount":46,"integrity":"sha512-rktEJfmw+gpQsLtDCqlZIZx9J1ViTbeUdZOIskIpNEhNDQ9pWr+ufaftZI76pWGkUoCzshvcCo1y7cNvdgzyfw==","signatures":[{"sig":"MEQCIFsmAXusHHl4kCMfFz3ZVVpzx7rBiuHATcwgH/7Lf4ASAiAAmBk3epBZh1PBNhoroCNG+jEmx3jh2zcHLwL+1xiXtQ==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":244926},"type":"module","gitHead":"5c98f142383cc66455fdd1d542df294f920d7353","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"12.0.2","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.20.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^5.0.0","@types/node":"^26.4.1","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-ai":">=0.84.0","@earendil-works/pi-tui":">=0.84.0","@earendil-works/pi-coding-agent":">=0.84.0"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.11.0_1788732118876_0.40334267306827654","host":"s3://npm-registry-packages-npm-production"}},"0.12.0":{"name":"pi-llama-cpp","version":"0.12.0","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.12.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"d47f7d386441ac5716e1df4a2bde5e81fc66df9e","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.12.0.tgz","fileCount":52,"integrity":"sha512-YPa0YgfdgZDjMfLSoJK3FKXruUn7nRIWBvqy69gDvZIHv/YBbTMezQ2gTDiYpNALcI6wuzAWz5DhopsOqv5tww==","signatures":[{"sig":"MEUCIQDNRev0EhvFloWtrpNpV5nf+Pv31QQnt9GdBvsCxaDaZQIgVZE1QaQlPAmQrr9Uvl7cFmG2/YhBUZf6OV1ApCDlETI=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"},{"sig":"MEUCIQDAjTs0G17Zllx0Zo5qqHhnt7YRENl7bRTN4PY474lKRwIgTixYNifBlgpbjQWngOtHrmwawwl5XQQ2323C+rp7rOU=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":314294},"type":"module","gitHead":"664a9f5ac9330ff626bc1fb04771af4dc1440310","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"12.0.2","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.20.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^5.0.0","@types/node":"^26.4.1","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-ai":">=0.85.1","@earendil-works/pi-tui":">=0.85.1","@earendil-works/pi-coding-agent":">=0.85.1"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.12.0_1789234334624_0.25679425812129697","host":"s3://npm-registry-packages-npm-production"}},"0.13.0":{"name":"pi-llama-cpp","version":"0.13.0","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.13.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"df9af88fee60ee33f46373d6e096849757f2ad0d","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.13.0.tgz","fileCount":54,"integrity":"sha512-hentHdxf87MgYs7BO5h1osnXlj1zlRRTZ0Ua5q26RzEpLOOAlqDUqtyqB71ZkXOmK86+kQAolSkkJRwPd0h4hw==","signatures":[{"sig":"MEQCICs96XwuWjG6tGu98bvUCnJFSEcsjflBXt65QDFwSXogAiAFuk3ESE0v8nIeYrjOFSrbshunUcXHvgVW3KTuzzPavA==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"},{"sig":"MEUCIQDlwb5mTDswnVYYMN+3PS12mgHuCRxgDVfg9761aXh1sgIgOVw5BBCWoJePaafGmeQlfp3CgVNOPAlgAKjFLf8UxRE=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":330497},"type":"module","gitHead":"814a47f57667af09437ec98ec5dd6d1d1dd53f51","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"12.0.2","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.21.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^5.0.1","@types/node":"^26.4.1","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-ai":">=0.85.1","@earendil-works/pi-tui":">=0.85.1","@earendil-works/pi-coding-agent":">=0.85.1"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.13.0_1789509281623_0.7544011052999613","host":"s3://npm-registry-packages-npm-production"}},"0.14.0":{"name":"pi-llama-cpp","version":"0.14.0","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.14.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"76037ecd60261af1e44f0ac982b2ea28b07e1676","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.14.0.tgz","fileCount":78,"integrity":"sha512-hFasUPoabODJpVIiONTmw1PjVuNA3A4LfbZWqVoAoNEuePcRXw2CmZFfmoZDdydyMV6JQEAshBhF06LEZFxG/w==","signatures":[{"sig":"MEQCICcven+LNa4xOiJvkoyWIbRiwRkdgSl0+1P8RjTqcxo5AiBDI+KQm92n688ghUP+MaYRRHrqm4FqAxY13ovhWCW1Jg==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"},{"sig":"MEUCICudgGq8QjyLh5lNWRokWpizUTLSONuyibAqnQUsX4x4AiEApSQLcAUD5YmItDTtNnNODO88eQjDLlXS6Ue9UJZxhyI=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":362248},"type":"module","gitHead":"503d1a11047aa986417388db1df237355bcd8b74","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"12.0.2","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.21.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^5.0.1","@types/node":"^26.6.2","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-ai":">=0.85.1","@earendil-works/pi-tui":">=0.85.1","@earendil-works/pi-coding-agent":">=0.85.1"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.14.0_1789865903852_0.5622415133067762","host":"s3://npm-registry-packages-npm-production"}},"0.15.0":{"name":"pi-llama-cpp","version":"0.15.0","keywords":["pi","pi-package","pi-extension","llama-cpp"],"author":{"name":"Gabriel Sanhueza"},"license":"MIT","_id":"pi-llama-cpp@0.15.0","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"pi":{"extensions":["./src/index.ts"]},"dist":{"shasum":"aeb9b635a4663d222a5d263e707f03b3dc34eaa8","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.15.0.tgz","fileCount":82,"integrity":"sha512-5kJt4SrU9eLmf6AW8SYgUh8i3MJ+phbLejwU/UjY8E9KwevNEEoniPrafIOqdrQUHXIFLtQnISyivcniT8pVfQ==","signatures":[{"sig":"MEQCIDqJhqvVoLI/aPrVfORZcNo/XoCQirXbyKV6m9txPiA3AiAtlOTpat28pNNOV9UXWXyQM/asdyQ8hbZoJ1b+GTw40g==","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"},{"sig":"MEUCIQDaY1t15XAxz7NaE6qYxVuLAuDo8PUpYBQUKiTbV2XR/gIgHGWs4QYG2AD48VbmGFJ0OP4VHr+e6W+6il77FEaodgg=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"}],"unpackedSize":369322},"type":"module","gitHead":"60afb32723ff1d3bc48a716af3842d6d424bdd65","scripts":{"test":"vitest run"},"_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"12.1.0","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"_nodeVersion":"24.21.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^5.0.2","@types/node":"^26.6.3","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-ai":">=0.85.1","@earendil-works/pi-tui":">=0.85.1","@earendil-works/pi-coding-agent":">=0.85.1"},"_npmOperationalInternal":{"tmp":"tmp/pi-llama-cpp_0.15.0_1790381188662_0.7957502705470876","host":"s3://npm-registry-packages-npm-production"}},"0.16.0":{"pi":{"extensions":["./src/index.ts"]},"_id":"pi-llama-cpp@0.16.0","bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"dist":{"shasum":"43644c57ce6850220e97fbb2bb6e10857414d866","tarball":"https://registry.npmjs.org/pi-llama-cpp/-/pi-llama-cpp-0.16.0.tgz","fileCount":87,"integrity":"sha512-81BRAZY+HOEH13J0C7i7KReHjckmGEQYr6GxvCcWRJVADnlTeFshOM9gKSSvUADFIEK8BrMbK20FkkLl4A5Gcw==","signatures":[{"sig":"MEUCIATK/qF5kFS4tJ2Ps3ZeRzZZgg7kUvYQso+FCod+5KbdAiEA75PP2j1plHI5BJrSLJ7uC57rEagE4ztxEh69LnCueyg=","keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U"},{"keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U","sig":"MEQCIFoOK/8CeXULaDjmz8GFRqXgv1pfqih5l6uSyjvUeoTIAiBxA+4zuK0OuoaysVVyM9UBU93wYSdDjOsGJAzmT8N03w=="}],"unpackedSize":391282},"name":"pi-llama-cpp","type":"module","author":{"name":"Gabriel Sanhueza"},"gitHead":"039e1c352c5f0be71e81965ce236e8936f44aa7d","license":"MIT","scripts":{"test":"vitest run"},"version":"0.16.0","_npmUser":{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"},"homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","keywords":["pi","pi-package","pi-extension","llama-cpp"],"prettier":{"plugins":["prettier-plugin-organize-imports"]},"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"_npmVersion":"12.1.0","description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","directories":{},"maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"_nodeVersion":"24.21.0","_hasShrinkwrap":false,"devDependencies":{"vitest":"^5.0.2","@types/node":"^26.6.3","prettier-plugin-organize-imports":"^4.3.0"},"peerDependencies":{"@earendil-works/pi-ai":">=0.85.1","@earendil-works/pi-tui":">=0.85.1","@earendil-works/pi-coding-agent":">=0.85.1"},"_npmOperationalInternal":{"host":"s3://npm-registry-packages-npm-production","tmp":"tmp/pi-llama-cpp_0.16.0_1790719175344_0.8350623655277958"}}},"time":{"created":"2026-04-27T00:36:59.124Z","modified":"2026-09-29T21:59:35.638Z","0.1.0":"2026-04-27T00:36:59.302Z","0.1.1":"2026-04-27T00:42:01.076Z","0.1.2":"2026-04-27T01:30:56.515Z","0.2.0":"2026-05-01T02:00:25.746Z","0.2.1":"2026-05-01T03:38:37.546Z","0.2.2":"2026-05-01T21:06:51.016Z","0.2.3":"2026-05-02T03:43:44.573Z","0.3.0":"2026-05-09T18:27:25.681Z","0.3.1":"2026-05-09T20:03:13.372Z","0.3.2":"2026-05-10T04:07:50.858Z","0.3.3":"2026-05-10T07:18:23.701Z","0.3.4":"2026-05-10T07:39:27.372Z","0.4.0":"2026-05-16T05:58:58.232Z","0.5.0":"2026-05-23T04:30:46.819Z","0.5.1":"2026-05-31T19:03:37.266Z","0.6.0":"2026-06-06T17:53:48.401Z","0.7.0":"2026-06-13T06:15:01.722Z","0.7.1":"2026-06-13T06:32:51.688Z","0.7.2":"2026-06-20T04:43:38.330Z","0.8.0":"2026-06-28T04:21:41.334Z","0.8.1":"2026-06-28T04:51:14.583Z","0.8.2":"2026-07-16T18:16:47.619Z","0.9.0":"2026-07-16T21:23:34.226Z","0.9.1":"2026-07-21T00:48:06.189Z","0.9.2":"2026-08-20T22:16:27.608Z","0.10.0":"2026-08-29T22:17:02.009Z","0.11.0":"2026-09-06T22:01:59.028Z","0.12.0":"2026-09-12T17:32:14.736Z","0.13.0":"2026-09-15T21:54:41.717Z","0.14.0":"2026-09-20T00:58:23.980Z","0.15.0":"2026-09-26T00:06:28.778Z","0.16.0":"2026-09-29T21:59:35.436Z"},"bugs":{"url":"https://github.com/gsanhueza/pi-llama-cpp/issues"},"author":{"name":"Gabriel Sanhueza"},"license":"MIT","homepage":"https://github.com/gsanhueza/pi-llama-cpp#readme","keywords":["pi","pi-package","pi-extension","llama-cpp"],"repository":{"url":"git+https://github.com/gsanhueza/pi-llama-cpp.git","type":"git"},"description":"Pi extension for llama.cpp integration. Supports router, single and legacy models. Supports multiple servers.","maintainers":[{"name":"gsanhueza","email":"gabriel.sanhueza@ing.uchile.cl"}],"readme":"# pi-llama-cpp\n\nA [Pi Coding Agent](https://pi.dev/) extension that integrates with running [llama.cpp servers](https://github.com/ggml-org/llama.cpp) to provide live model browsing, loading, and switching directly from Pi.\n\n## Features\n\n- **Auto-detect models** — discovers all models available on your running llama.cpp server\n- **Live status indicators** — see which models are loaded, loading, failed, sleeping, or unloaded with color-coded icons\n- **Load / unload / switch** — manage models directly from the Pi command palette\n- **Multi-model router support** — works with both single-model and multi-model llama.cpp server configurations\n- **Image capabilities detection** — detects multimodal models automatically\n- **Flexible URL resolution** — configures the server via `llamaSettings` (project/global), environment variable, or legacy `llamaServerUrl`\n- **Auth support** — allows to login into a llama.cpp server that was secured with an API key\n- **Multiple server support** — connect to multiple llama.cpp servers simultaneously via `llamaSettings.servers` or semicolon-separated URLs\n- **Basic llama-swap support** — auto-detects and provides basic integration with [llama-swap](https://github.com/mostlygeek/llama-swap) gateways\n- **Thinking budget support** — configurable token budgets for model reasoning/thinking, mapped to Pi's thinking levels\n- **Real-time progress tracking** — live loading progress via SSE (falls back to polling)\n\n### Status Indicators\n\n| Icon | Status | Description |\n| ---- | -------- | --------------------------------- |\n| 🟢 | Loaded | Model is active and ready to use |\n| 🟡 | Loading | Model is currently being loaded |\n| 🔴 | Failed | Model failed to load |\n| 🔵 | Sleeping | Model is available, but inactive |\n| ⚪ | Unloaded | Model is not loaded on the server |\n\n> **Note**: The `Sleeping` status only shows when you start your server with `llama-server --sleep-idle-seconds ...`.\n> This is a **llama.cpp server flag** that tells the server to put idle models to sleep after `n` seconds.\n> The model awakens automatically when you send a message.\n\n> **Note:** You can run your server with API authentication with `llama-server --api-key ...`.\n\n### Server Health Indicators\n\nWhen browsing servers via `/models servers`, each server URL is prefixed with a health indicator:\n\n| Icon | Status | Description |\n| ---- | ------------ | --------------------------------------------- |\n| 🟢 | Healthy | Server responded successfully to health check |\n| 🟡 | Timeout | Server health check timed out |\n| 🔴 | Unreachable | Server could not be reached |\n| ⛔ | Unauthorized | Server requires an API key |\n\n## Installation\n\nThis package is a Pi extension. Install it with\n\n```bash\npi install npm:pi-llama-cpp\n```\n\nor\n\n```bash\npi install https://github.com/gsanhueza/pi-llama-cpp\n```\n\n## Configuration\n\nThe extension resolves the llama.cpp server configuration using the following priority order:\n\n1. **Environment variable** — `LLAMA_SERVER_URL`\n2. **`llamaSettings`** — Main configuration format in `.pi/settings.json` (project) or `~/.pi/agent/settings.json` (global)\n3. **`llamaServerUrl`** — Legacy format in `.pi/settings.json` (project) or `~/.pi/agent/settings.json` (global)\n4. **Default** — `http://127.0.0.1:8080`\n\n### Server configuration\n\nThe recommended way to configure the extension is using the `llamaSettings` key. This provides a structured way to define multiple servers with custom names and IDs, plus additional behavior options.\n\nAdd this to your `.pi/settings.json` (project) or `~/.pi/agent/settings.json` (global):\n\n#### Minimal configuration\n\n```json\n{\n \"llamaSettings\": {\n \"servers\": [{ \"url\": \"http://127.0.0.1:8080\" }]\n }\n}\n```\n\n#### Full configuration\n\n```json\n{\n \"llamaSettings\": {\n \"servers\": [\n {\n \"url\": \"http://127.0.0.1:8080\",\n \"id\": \"local\",\n \"name\": \"Local Server\",\n \"overrides\": {}\n },\n {\n \"url\": \"http://10.0.0.5:8080\",\n \"name\": \"Remote Server\"\n }\n ],\n \"reactToModelSelect\": true,\n \"autoloadOnMessage\": false,\n \"sortBy\": \"asc\",\n \"pollingTimeout\": 60000,\n \"serverTimeout\": 1000\n }\n}\n```\n\nWith this config, the servers will appear in Pi as **Llama.cpp (Local Server)** and **Llama.cpp (Remote Server)**.\n\n#### Server Options\n\n| Option | Type | Required | Description |\n| ------ | ------ | -------- | ---------------------------------------------------------------------------- |\n| `url` | string | Yes | The URL of the llama.cpp server |\n| `id` | string | No | Custom provider ID (used for API key auth). Defaults to `llama-server=` |\n| `name` | string | No | Display name for the server in the UI (shown as `Llama.cpp ()`) |\n\n> **Note:** If you set a custom `id`, you can use it in `~/.pi/agent/auth.json`. The extension will also fall back to the URL-based ID if no key is found for the custom `id`.\n\n#### Settings Options\n\n| Option | Type | Default | Description |\n| -------------------- | ------- | ------- | ------------------------------------------------------------- |\n| `reactToModelSelect` | boolean | `true` | Load the model when you switch via Pi's model picker. |\n| `autoloadOnMessage` | boolean | `false` | Automatically load an unloaded model before sending a message |\n| `sortBy` | string | `\"asc\"` | Sort order for models (see below) |\n| `pollingTimeout` | number | `60000` | Max time (ms) to wait for model loading before giving up |\n| `serverTimeout` | number | `1000` | Timeout (ms) for server health checks and SSE probes |\n| `showServerUrls` | boolean | `true` | Show `[Server: ]` suffix in the /models model list |\n\n> **Note:** `serverTimeout` controls individual HTTP request timeouts (health checks, SSE probe). `pollingTimeout` controls the total wait time for a model to finish loading. Increase `serverTimeout` for slow/high-latency servers, and `pollingTimeout` for large models or slow hardware.\n\n#### In-session settings menu\n\nRun `/models settings` to edit the scalar settings above without hand-editing JSON. Changes are written to the **project** `.pi/settings.json` if it exists, otherwise to **global** `~/.pi/agent/settings.json`. Boolean and sort changes apply immediately; timeout changes apply on the next model load. The `servers` list is edited with `/models servers` (see below), and per-server model overrides with `/models overrides` (see [Model Overrides](#model-overrides)).\n\n#### Server list editor\n\nRun `/models servers` to add, edit or remove entries of `llamaSettings.servers`\nwithout hand-editing JSON. Each server URL is prefixed with a status indicator\n(🟢 healthy, 🟡 timeout, 🔴 unreachable, ⛔ unauthorized) that reflects the\nresult of a health check and an auth probe against the server. Each change is\nwritten immediately to the **project** `.pi/settings.json` if it exists,\notherwise to **global** `~/.pi/agent/settings.json`.\n\nChanges take effect immediately after closing the editor: new servers\nregister their providers, removed ones leave pi's registry right away,\nand edited ones are re-registered with the fresh config — no restart or\n`/models` needed.\n\nLimitation: a model already loading in the background on a removed or\nedited server finishes loading, but its progress notifications stop;\nre-select it from the (new) provider afterwards.\n\n#### Environment variable\n\nFor a quick setup, you can use the `LLAMA_SERVER_URL` environment variable instead of the JSON config:\n\n```bash\nexport LLAMA_SERVER_URL=\"http://127.0.0.1:8080\"\n```\n\nThis is equivalent to defining a single server in `llamaSettings.servers` with just a URL.\n\n### Legacy configuration\n\nFor a simpler setup, you can use the legacy `llamaServerUrl` key:\n\n```json\n{\n \"llamaServerUrl\": \"http://127.0.0.1:8080\"\n}\n```\n\nThis is equivalent to defining a single server in `llamaSettings.servers` with just a URL.\n\n### Multiple servers\n\nTo connect to multiple llama.cpp servers simultaneously:\n\n**Using `llamaSettings` (recommended):**\n\n```json\n{\n \"llamaSettings\": {\n \"servers\": [\n { \"url\": \"http://127.0.0.1:8080\" },\n { \"url\": \"http://127.0.0.1:8081\" },\n { \"url\": \"http://10.0.0.5:8080\" }\n ]\n }\n}\n```\n\n**Using the environment variable:**\n\n```bash\nLLAMA_SERVER_URL=\"http://127.0.0.1:8080;http://127.0.0.1:8081;http://10.0.0.5:8080\"\n```\n\nEach server gets its own provider (e.g., **Llama.cpp (http://127.0.0.1:8080)**) and its own set of models. The `/models` command lists all models from all servers, labeled with their server URL.\n\n### API Key\n\nIf your llama.cpp server requires authentication, use `/login` in Pi, select the \"API key\" option, and choose the provider from the list that correlates with the server needing the API key.\n\nAlternatively, configure the API key in `~/.pi/agent/auth.json`:\nUse the provider ID `llama-server=` (or your custom `id` if you set one in `llamaSettings.servers`).\n\nThe `key` field supports several formats:\n\n| Format | Example | Description |\n| ----------------- | -------------------------------------------- | ------------------------------------------ |\n| **Literal** | `\"sk-abc123\"` | API key stored directly |\n| **Env ref** | `\"$OPENAI_API_KEY\"` or `\"${OPENAI_API_KEY}\"` | Resolved from `process.env` or `env` field |\n| **Shell command** | `\"!cat ~/.secrets/llama-key\"` | Stdout of the command is used |\n| **Escape** | `\"$$literal\"` | `$$` → literal `$`, `$!` → literal `!` |\n\n```json\n{\n \"llama-server=http://127.0.0.1:8080\": {\n \"type\": \"api_key\",\n \"key\": \"sk-abc123\"\n },\n \"llama-server=https://some-url-for-llama-cpp\": {\n \"type\": \"api_key\",\n \"key\": \"$LLAMA_API_KEY\"\n },\n \"llama-server=https://secure-server\": {\n \"type\": \"api_key\",\n \"key\": \"!cat ~/.secrets/llama-key\"\n },\n \"llama-server=https://braced-ref\": {\n \"type\": \"api_key\",\n \"key\": \"${API_KEY}\"\n }\n}\n```\n\nFor env ref formats, you can also store the variable value alongside the key using the `env` field:\n\n```json\n{\n \"llama-server=http://127.0.0.1:8080\": {\n \"type\": \"api_key\",\n \"key\": \"$MY_KEY\",\n \"env\": {\n \"MY_KEY\": \"sk-abc123\"\n }\n }\n}\n```\n\n## Usage\n\n### Prerequisites\n\nMake sure your llama.cpp server is running with the appropriate flags.\n\n- For multi-model support (model router), start the server with:\n\n```bash\nllama-server --models-preset path/to/presets.ini ...\n```\n\n- For single-model mode, start the server with:\n\n```bash\nllama-server --model path/to/model.gguf ...\n```\n\n- For legacy-model mode (e.g., [ik_llama.cpp](https://github.com/ikawrakow/ik_llama.cpp)), the extension auto-detects and handles it transparently.\n\n> **Note:** This extension is focused on llama.cpp, not on ik_llama.cpp. Nonetheless, since I found a way to make it work with this extension, I added the option.\n\n> **Note:** The ik_llama.cpp fork is not legacy at all, but it uses an old way of describing models compared to llama.cpp.\n\n- For llama-swap mode, point the extension at a running [llama-swap](https://github.com/mostlygeek/llama-swap) instance instead of a raw llama.cpp server. The extension auto-detects this mode via the `src` field in the server props response.\n\n> **Note:** llama-swap support is basic — only model listing, status, load/unload, and capability detection are implemented.\n\nThe extension determines the context size as follows:\n\n- A per-model `contextSize` override (see [Model Overrides](#model-overrides)) takes precedence over everything below\n- **Router mode**\n - When loaded, reads `meta.n_ctx` from the `/v1/models` endpoint\n - When not loaded, reads `--ctx-size` and/or `--fit-ctx` from the server arguments (which can also originate from the **presets.ini** file the llama.cpp server uses to load its models).\n- **Single mode** — reads `meta.n_ctx` from the `/v1/models` endpoint\n- **Legacy mode** — reads `max_model_len` from `/v1/models`, falling back to `n_ctx` from `/props`\n- **Llama-swap mode** — reads `meta.n_ctx` from the llama-swap server via `/v1/models`\n- Falls back to `128000` if not available\n\n### Commands\n\n| Command | Description |\n| ------------------- | --------------------------------------------------------------------------------------- |\n| `/models` | Browse your models with live status. Select a model to load, switch, or unload it. |\n| `/models info` | Show detailed information for all available models at once. |\n| `/models unload` | Unload all loaded models at once. |\n| `/models settings` | Open a menu to edit the scalar `llamaSettings` fields. |\n| `/models servers` | Add, edit or remove llama.cpp server URLs via a TUI editor. |\n| `/models overrides` | Edit per-server model overrides (`llamaSettings.servers[].overrides`) via a TUI editor. |\n\n> **Note:** When a llama.cpp server is slow to respond, it will be skipped at startup with a warning. Run `/models` to retry without timeout and see all models.\n\n> **Note:** When a llama.cpp server is unreachable, `/models` displays an error notification with the configured server URL, but healthy servers continue to show their models.\n\n> **Note:** The `/models unload` command only makes sense in router mode.\n\n#### Model sorting\n\nThe order of models in the `/models` menu is controlled by the `sortBy` setting.\nServers maintain their order from `llamaSettings`; sorting applies **within each server**:\n\n| Value | Description |\n| ------------- | ------------------------------------------------------------------------------------------------------------------------- |\n| `\"asc\"` | Sort by model ID ascending (default) |\n| `\"desc\"` | Sort by model ID descending |\n| `\"asc-name\"` | Sort by model name ascending (ties broken by ID) |\n| `\"desc-name\"` | Sort by model name descending (ties broken by ID) |\n| `\"api\"` | No sorting — models appear in the order returned by each server's `/v1/models` endpoint, servers in `llamaSettings` order |\n\n### Model Actions\n\nWhen browsing models via the `/models` command, you can:\n\n- **Load & switch** — Load an unloaded model and switch to it\n- **Switch model** — Switch to a model that is already loaded\n- **Unload** — Unload a loaded model to free memory\n- **Retry** — Retry loading a failed model\n- **Info** — View model details (ID, capabilities, context size)\n- **Cancel** — Cancel the current operation\n\n> **Note:** In single-model and legacy-model mode, **Unload** is not available, since there is only one model on the server.\n\n### Thinking Budgets\n\nThe extension supports configurable **thinking budgets** that control how many tokens the model allocates to its reasoning/thinking process.\nThis is tied to Pi's thinking level selector (off, minimal, low, medium, high, xhigh, max).\n\n| Level | Tokens | Description |\n| --------- | ------ | ---------------------------- |\n| `off` | 0 | Thinking disabled |\n| `minimal` | 1,024 | Short reasoning steps |\n| `low` | 2,048 | Light reasoning |\n| `medium` | 8,192 | Balanced reasoning (default) |\n| `high` | 16,384 | Extended reasoning |\n| `xhigh` | 32,768 | Deep reasoning |\n| `max` | -1 | Unlimited reasoning |\n\nUser-defined budgets can override the defaults by adding a `thinkingBudgets` object to `~/.pi/agent/settings.json` (global) or `.pi/settings.json` (per-project):\n\n```json\n{\n \"thinkingBudgets\": {\n \"minimal\": 256,\n \"low\": 1024,\n \"medium\": 2048,\n \"high\": 4096,\n \"xhigh\": 8192\n }\n}\n```\n\nOnly `minimal`, `low`, `medium`, `high` and `xhigh` are configurable — `off` (0) and `max` (-1, unlimited) are fixed.\nThe extension automatically injects the appropriate `thinking_budget_tokens` into each request payload based on the selected level.\n\n### Model Overrides\n\nA locally-run `llama.cpp` server is free, but you can simulate costs for budgeting, experimentation, or comparison purposes — and fine-tune what the extension reports about each model.\n\nThis extension supports **per-model, per-server configuration** via the `overrides` key inside each server entry of `llamaSettings.servers`. Each entry can override the model's `cost`, `capabilities`, `reasoning`, `contextSize`, `maxTokens`, and `compat`, regardless of what the server reports.\n\nAdd overrides to your server configuration:\n\n```json\n{\n \"llamaSettings\": {\n \"servers\": [\n {\n \"url\": \"http://127.0.0.1:8080\",\n \"overrides\": {\n \"qwen-3.8-27b\": {\n \"cost\": { \"input\": 0.42, \"output\": 3.0, \"cacheRead\": 0.085 }\n },\n \"glm-5.3-flash\": {\n \"cost\": { \"input\": 0.15, \"output\": 0.5, \"cacheRead\": 0.03 },\n \"capabilities\": [\"text\"],\n \"reasoning\": false,\n \"contextSize\": 32768,\n \"maxTokens\": 4096\n }\n }\n }\n ]\n }\n}\n```\n\nEvery field of an override is optional — absent fields fall back to what the extension detects (`capabilities`) or to its defaults (`reasoning: true`, zeroed cost).\n\n#### Override editor\n\nRun `/models overrides` to edit a server's override entries without hand-editing\nJSON. It opens a settings menu (same UX as `/models settings`).\n\nEach change is written immediately to the **project**\n`.pi/settings.json` if it exists, otherwise to **global**\n`~/.pi/agent/settings.json`.\n\nOverrides take effect on the next provider request after closing the editor —\nno `/reload` needed.\n\n#### Cost Fields\n\nInside an override, the `cost` object accepts:\n\n| Field | Type | Description |\n| ------------ | ------ | ----------------------------------- |\n| `input` | number | Cost per million input tokens |\n| `output` | number | Cost per million output tokens |\n| `cacheRead` | number | Cost per million cache read tokens |\n| `cacheWrite` | number | Cost per million cache write tokens |\n\nAll four fields are optional — unspecified fields default to zero.\nIn the override editor, entering `0` (or leaving a field empty) removes the field from the settings — and the `cost` object itself once no fields remain — with the same effect as leaving it unset.\n\n#### Other Fields\n\n| Field | Type | Description |\n| -------------- | ---------------- | ---------------------------------------------------------------------------------------------- |\n| `capabilities` | array of strings | Pi capabilities for the model (`\"text\"`, `\"image\"`). Fully replaces the detected capabilities. |\n| `reasoning` | boolean | Whether the model is a reasoning model. Defaults to `true` when absent. |\n| `contextSize` | number | Override the model's context size in tokens. Falls back to autodetection when absent or `0`. |\n| `maxTokens` | number | Override max generation tokens. Falls back to context size when absent or `0`. |\n| `compat` | object | OpenAI-compatible provider compatibility settings (see below). |\n\n#### Compatibility (`compat`)\n\nThe `compat` field accepts any subset of [OpenAI-compatible provider compatibility settings](https://github.com/earendil-works/pi/blob/main/packages/ai/src/types.ts) used by the `openai-completions` API. These control how the extension talks to your server — for example, disabling `developer` role support, choosing the thinking format, enabling Anthropic-style cache control, or setting thinking token budgets.\n\nExample:\n\n```json\n{\n \"llamaSettings\": {\n \"servers\": [\n {\n \"url\": \"http://127.0.0.1:8080\",\n \"overrides\": {\n \"llama-3\": {\n \"compat\": {\n \"supportsDeveloperRole\": false,\n \"thinkingFormat\": \"openai\",\n \"thinkingTokenBudgetField\": \"thinking_budget_tokens\"\n }\n }\n }\n }\n ]\n }\n}\n```\n\n### Prefix matching\n\nOverride keys are treated as **prefix filters** — a model ID matches if it starts with the key. When multiple patterns match, the **longest (most specific) match wins**. This lets you define broad patterns at the top of your overrides and override them with more specific ones below.\n\nExample:\n\n```json\n{\n \"llama\": { \"cost\": { \"input\": 0.01, \"output\": 0.02 } },\n \"llama-3\": { \"reasoning\": false },\n \"llama-3-8b\": { \"cost\": { \"input\": 0.2, \"output\": 0.6 } }\n}\n```\n\n| Model ID | Matching keys | Winner (longest) | Effective override |\n| ------------- | -------------------------------- | ---------------- | ------------------------------------------------------ |\n| `llama-3-8b` | `llama`, `llama-3`, `llama-3-8b` | `llama-3-8b` | `{ cost: { input: 0.2, output: 0.6 } }` |\n| `llama-3-70b` | `llama`, `llama-3` | `llama-3` | `{ reasoning: false }` |\n| `mistral-7b` | `llama` (no) | none | defaults (zero cost, detected caps, `reasoning: true`) |\n\n> **Note:** Exact model IDs still work — they are simply the longest possible prefix for themselves. Empty keys are silently ignored.\n\nModel matching uses this prefix system — the model ID must start with the override key for a match.\n\n> **Note:** Overrides are resolved through the same settings merge logic (project overrides global), so they follow the same precedence chain as other server settings. If the same URL appears multiple times with different `overrides`, only the first one's overrides will be used (consistent with existing dedup behavior).\n\n### Model Selection Event\n\nWhen you switch models via Pi's model picker (instead of using the `/models` command), the extension listens for the `model_select` event, which also loads the requested model before the conversation begins.\n\nThis keeps the server in sync with the active model in Pi, regardless of how the switch was initiated — you don't need to manually load models before using them.\n\nYou can disable this behavior by setting `reactToModelSelect` to `false` in `llamaSettings`.\n\n> **Note:** If you switch sessions while a model load is in-flight, you'll see a warning, but the load continues in the background. Use `/models` in the new session to verify the model status.\n\n### Loading Models\n\nWhen you trigger a load, switch, or retry action, the extension uses SSE (Server-Sent Events) to receive real-time progress updates from the server. If SSE is not available, it falls back to polling.\n\nIf loading takes longer than **60 seconds** (configurable via `pollingTimeout`), the operation times out with an error.\n\n> **Note:** The timeout only applies to the progress detection. The model might still be loading in the background.\n\n### Model Configuration\n\nEach model exposed to Pi includes the following defaults:\n\n- **`contextWindow`** — detected from llama-server (see how the extension determines the context size above); can be overridden per-model via `llamaSettings.servers[].overrides` (see [Model Overrides](#model-overrides))\n- **`maxTokens`** — dynamically set to the model's context window (detected from llama-server); can be overridden per-model via `llamaSettings.servers[].overrides` (see [Model Overrides](#model-overrides))\n- **`reasoning`** — `true` by default (llama.cpp's `/v1/models` endpoint does not expose it); can be overridden per-model via `llamaSettings.servers[].overrides` (see [Model Overrides](#model-overrides))\n- **`cost`** — all zero by default; can be customized per-model via `llamaSettings.servers[].overrides` (see [Model Overrides](#model-overrides))\n- **`compat`** — OpenAI-compatible provider compatibility settings; can be set per-model via `llamaSettings.servers[].overrides` (see [Model Overrides](#model-overrides))\n\n## Dependencies\n\n| Peer dependency | Purpose |\n| --------------------------------- | ------------------- |\n| `@earendil-works/pi-ai` | Pi AI SDK |\n| `@earendil-works/pi-coding-agent` | Pi Coding Agent SDK |\n| `@earendil-works/pi-tui` | Pi TUI SDK |\n","readmeFilename":"README.md"}