Edits

docs: render site edits

2026-06-02 · 703ca0d

index.html

-Linea
 Local AI chat.
 A Go server. A React chat UI. PostgreSQL when configured. Model fallback when needed.
 Why Linea is local-first
 1 Jun 2026
+Edits
+2 Jun 2026

docs: simplify site copy (#42)

2026-06-02 · c061497

index.html

-Notes
-Local tools should stay understandable.
-Linea is a local-first AI assistant. These notes record design choices, limits, and small decisions that keep the tool clear.
+Linea
+Local AI chat.
+A Go server. A React chat UI. PostgreSQL when configured. Model fallback when needed.
 Why Linea is local-first
 1 Jun 2026

notes/local-first.html

 1 Jun 2026
-Why Linea is local-first
-A local assistant should start with the machine in front of the user.
-The interface is served by one Go process. The web UI is the client. Conversation data stays in PostgreSQL when a database is configured, and memory storage remains available for a smaller setup.
-Model access is external when needed. Gemini is the primary provider. Cerebras and SambaNova use OpenAI-compatible endpoints. Ollama gives a local fallback when remote quota or traffic gets in the way.
-This shape keeps the system inspectable. The user can run the server, check the health endpoint, read the API, and replace parts without learning a larger platform.
-Local-first does not mean offline-only. It means the tool has a clear center. The local server owns the app. Providers are replaceable. The interface stays small.
+Local-first
+Linea starts on the user machine.
+One Go process serves the API and the web UI. PostgreSQL stores conversations when configured. Memory storage works without a database.
+Gemini is primary. Cerebras and SambaNova use OpenAI-compatible endpoints. Ollama is the local fallback.
+The app can be checked from the command line. Run the server. Check `/healthz`. Read the API. Replace a provider.
+Local-first does not mean offline-only. It means the local server is the center.

build: package native macos window

2026-06-01 · 84f2055

notes/local-first.html

 Why Linea is local-first
 A local assistant should start with the machine in front of the user.
-The interface is served by one Go process. The browser is the client. Conversation data stays in PostgreSQL when a database is configured, and memory storage remains available for a smaller setup.
+The interface is served by one Go process. The web UI is the client. Conversation data stays in PostgreSQL when a database is configured, and memory storage remains available for a smaller setup.
 Model access is external when needed. Gemini is the primary provider. Cerebras and SambaNova use OpenAI-compatible endpoints. Ollama gives a local fallback when remote quota or traffic gets in the way.
 This shape keeps the system inspectable. The user can run the server, check the health endpoint, read the API, and replace parts without learning a larger platform.

docs: publish linea notes site

2026-06-01 · a10e651

index.html

+Notes
+Local tools should stay understandable.
+Linea is a local-first AI assistant. These notes record design choices, limits, and small decisions that keep the tool clear.
+Why Linea is local-first
+1 Jun 2026

notes/local-first.html

+1 Jun 2026
+Why Linea is local-first
+A local assistant should start with the machine in front of the user.
+The interface is served by one Go process. The browser is the client. Conversation data stays in PostgreSQL when a database is configured, and memory storage remains available for a smaller setup.
+Model access is external when needed. Gemini is the primary provider. Cerebras and SambaNova use OpenAI-compatible endpoints. Ollama gives a local fallback when remote quota or traffic gets in the way.
+This shape keeps the system inspectable. The user can run the server, check the health endpoint, read the API, and replace parts without learning a larger platform.
+Local-first does not mean offline-only. It means the tool has a clear center. The local server owns the app. Providers are replaceable. The interface stays small.