\n \n
\n\n\n**Swarmdo's web UI is a multi-model AI chat with built-in Model Context Protocol (MCP) tool calling.** Talk to Qwen, Claude, Gemini, or OpenAI while Swarmdo invokes the same MCP tools the CLI uses — agent orchestration, persistent memory, swarm coordination, code review, GitHub ops — directly from chat. No install, no API key needed to try it.\n\n| | What it is | Why it matters |\n|---|------------|----------------|\n| 🧠 | **Any model, local or remote** | 6 curated frontier models out-of-the-box — Qwen 3.6 Max (default), Claude Sonnet 4.6, Claude Haiku 4.5, Gemini 2.5 Pro, Gemini 2.5 Flash, OpenAI — via OpenRouter. Add your own: any OpenAI-compatible endpoint (vLLM, Ollama, LM Studio, Together, Groq, self-hosted). |\n| 🦾 | **swarmLLM self-learning AI** | Native support for [swarmLLM](the upstream project (see NOTICE)) (lives in `upstream/SwarmVector/examples/swarmLLM`) — Swarmdo's self-improving local model layer. Routes to MicroLoRA adapters, learns from your trajectories via SONA, and stays on your machine. Pair with the cloud models or run fully offline. |\n| 🛠️ | **~210 tools, ready to call** | 5 server groups (Core, Intelligence, Agents, Memory, DevTools) plus an 18-tool gallery that runs entirely in your browser — works offline. |\n| 🔌 | **Bring your own MCP servers** | Click the **MCP (n)** pill in the chat input → *Add Server* and paste any MCP endpoint (HTTP, SSE, or stdio). Your tools join Swarmdo's native ones in the same parallel-execution flow. Run a local MCP server on `localhost:3000` and it just works. |\n| ⚡ | **Tools run in parallel** | One model response can fire 4–6+ tools at the same time. The UI shows them as cards with a *Step 1 — 2 tools completed* badge so you can see exactly what ran. |\n| 💾 | **Memory that sticks** | Say *\"remember my favorite color is indigo\"* and ask weeks later — Swarmdo recalls it. Backed by AgentDB + HNSW vector search (measured ~1.9x–4.7x faster than brute force above the crossover, recall@10 ~0.99). |\n| 📘 | **Built-in capabilities tour** | Click the question-mark icon in the sidebar — a \"Swarmdo Capabilities\" modal opens with the full tool list, model strengths, architecture, and keyboard shortcuts. |\n| 🚀 | **Zero install to try** | Open the hosted URL, pick a model, type a question. That's the whole onboarding. |\n\n\n\n \n
\n\n\n\n| | What it is | Why it matters |\n|---|------------|----------------|\n| 🎯 | **Plain-English goals** | Type *\"ship the auth refactor with tests and a PR\"* — Swarmdo extracts the success criteria, the constraints, and the implicit preconditions. No JSON, no DSL. |\n| 🧭 | **GOAP A\\* planner** | Classic gaming-AI planning ported to software work: state-space search through actions with preconditions/effects to find the shortest viable path. Replans on the fly when state changes. |\n| 🌳 | **Visual plan tree** | Goals render as collapsible action trees with progress, blocked branches, and rollbacks highlighted. See *exactly* why an agent picked a path — no opaque chain-of-thought. |\n| ♻️ | **Adaptive replanning** | When an action fails or new info arrives, the planner re-runs A\\* from the current state instead of restarting. Failures become learning, not loops. |\n| 🧠 | **Shared memory + SONA** | Plans, trajectories, and outcomes flow into AgentDB. Future plans retrieve past solutions via HNSW — the planner gets smarter with every run. |\n| 🔗 | **Wired to MCP tools** | Every action node maps to a tool call (Swarmdo's ~210 MCP tools, your custom servers, or shell). The planner schedules them in parallel where the dependency graph allows. |\n\n\n### Agent Federation — Slack for Agents\n\n```\nYour Agent --> [ Remove secrets ] --> [ Sign message ] --> [ Encrypted channel ]\n Emails, SSNs, Proves it came No one reads it\n keys stripped from you in transit\n |\n v\nTheir Agent <-- [ Block attacks ] <-- [ Check identity ] <------+\n Stops prompt Rejects forgeries\n injection\n\n Audit trail on both sides.\n Trust builds over time. Bad behavior = instant downgrade.\n```\n\nSlack gave teams channels. Federation gives agents the same thing — **shared workspaces across trust boundaries**, where agents on different machines, orgs, or cloud regions can discover each other, prove who they are, and collaborate on tasks.\n\nThe difference: some channels are trusted, some aren't. [`@swarmdo/plugin-agent-federation`](the upstream project (see NOTICE)) handles that automatically. Your agents join a federation, get verified via mTLS + ed25519, and start exchanging work — with PII stripped before anything leaves your node and every message auditable. Untrusted agents can still participate at lower privilege: they see discovery info, not your memory. As they prove reliable, trust upgrades. If they misbehave, they get downgraded instantly — no human in the loop required.\n\nYou don't configure handshakes or manage certificates. You `federation init`, `federation join`, and your agents start talking. The protocol handles identity, the PII pipeline handles data safety, and the audit trail handles compliance.\n\n> **📘 Full user guide:** [`docs/federation/`](./docs/federation/) — setup, MCP tools, trust levels, circuit breaker, and the (opt-in) WireGuard mesh layer that ties packet-layer reachability to federation trust. ADR-111 deep-dive at [`docs/federation/phase7-mesh-bringup.md`](./docs/federation/phase7-mesh-bringup.md).\n\n