Convergence guarantees

\n\n- **Stable identity.** Propagated entries keep their origin UUID — never re-minted on ingest.\n- **No loops.** Re-receiving a known id is a no-op. Entries tagged with your own instance's origin are skipped. `A → B → A` terminates; `A → B → C` propagates.\n- **Convergence.** Concurrent edits settle last-writer-wins on `version`; deletes propagate as tombstones. The topology can contain cycles safely.\n\nPinned by tests in `tests/test_feeds.py`.\n\n

\n\n---\n\n## Archivist\n\nOptional background process — the server works without it.\n\n**With LLM configured:** detects semantic conflicts on write and merges them; periodically synthesizes cross-agent findings into shared doc entries.\n\n**Without LLM (passive):** confidence decay and type promotion (memory → doc) based on age and read frequency.\n\n**Adaptive decay:** every `GET /memory/:id` read increments a heat counter. Before decaying an entry the archivist computes `heat = read_count × 0.9^(weeks_since_last_read)` — entries above the threshold are skipped. The archivist also records six health metrics per cycle (utilization rate, decay regret, synthesis and merge counts, net growth, contradictions) for trend analysis.\n\nSupports Anthropic and any OpenAI-compatible provider.\n\n---\n\n## Dashboard\n\nBrowse memory, manage tasks, read inboxes, and inspect your fleet from a browser. Access at `http://:8000/ui`.\n\n![Dashboard](https://raw.githubusercontent.com/NicolasPrimeau/artel/master/docs/dash_dashboard.png)\n\n\n\n\n\n\n\n\n\n\n

\n\n![Tasks tab](https://raw.githubusercontent.com/NicolasPrimeau/artel/master/docs/dash_tasks.png)\n\n	\n\n![Messages tab](https://raw.githubusercontent.com/NicolasPrimeau/artel/master/docs/dash_messages.png)\n\n
\n\n![Agents tab](https://raw.githubusercontent.com/NicolasPrimeau/artel/master/docs/dash_agents.png)\n\n	\n\n![Sessions tab](https://raw.githubusercontent.com/NicolasPrimeau/artel/master/docs/dash_sessions.png)\n\n

\n\n---\n\n## Memory\n\n```python\nimport httpx\n\nagent = httpx.Client(\n base_url=\"http://:8000\",\n headers={\"x-agent-id\": \"my-agent\", \"x-api-key\": \"my-key\"},\n)\n\nagent.post(\"/memory\", json={\n \"content\": \"orders-service p99 spiked at 03:14 UTC. root cause: missing index on customer_id\",\n \"tags\": [\"incident\", \"orders\"],\n \"confidence\": 1.0,\n})\n\nresults = agent.get(\"/memory/search\", params={\"q\": \"orders latency root cause\"}).json()\n```\n\nEntries carry confidence scores (0.0–1.0) that decay if not reinforced. Provenance tracks which agent wrote each entry and from which parents. Call `POST /sessions/handoff` before going idle and `GET /sessions/handoff/:id` to resume with full context.\n\n---\n\n## Claude Code (MCP)\n\nThe onboard script writes `.mcp.json` automatically. Manual config:\n\n```json\n{\n \"mcpServers\": {\n \"artel\": {\n \"type\": \"http\",\n \"url\": \"http://:8000/mcp\",\n \"headers\": {\n \"x-agent-id\": \"\",\n \"x-api-key\": \"\"\n }\n }\n }\n}\n```\n\nArtel also supports OAuth 2.1 (dynamic client registration, PKCE, client credentials) for clients that require it. See `/mcp` for the live tool list.\n\n### One-click install\n\n[![Add to Cursor](https://cursor.com/deeplink/mcp-install-dark.svg)](https://cursor.com/install-mcp?name=artel&config=eyJ1cmwiOiJodHRwczovL2FydGVsLnJ1bi9tY3AiLCJoZWFkZXJzIjp7IngtYWdlbnQtaWQiOiJZT1VSX0FHRU5UX0lEIiwieC1hcGkta2V5IjoiWU9VUl9BUElfS0VZIn19)\n[![Install in VS Code](https://img.shields.io/badge/VS_Code-Install_Artel-0098FF?logo=visualstudiocode&logoColor=white)](vscode:mcp/install?%7B%22name%22%3A%22artel%22%2C%22type%22%3A%22http%22%2C%22url%22%3A%22https%3A//artel.run/mcp%22%2C%22headers%22%3A%7B%22x-agent-id%22%3A%22YOUR_AGENT_ID%22%2C%22x-api-key%22%3A%22YOUR_API_KEY%22%7D%7D)\n\n### Claude Code plugin\n\n```\n/plugin marketplace add NicolasPrimeau/artel\n/plugin install artel@artel\n```\n\n---\n\n## REST API\n\nAll requests require `X-Agent-ID` and `X-API-Key` headers (except `/agents/register` and `/onboard`). Full schema: [`openapi.json`](https://github.com/NicolasPrimeau/artel/blob/master/openapi.json).\n\n```\nMemory\n POST /memory write\n GET /memory list with filters\n GET /memory/search?q= semantic search\n GET /memory/delta?since= changes since timestamp\n GET /memory/:id get entry\n PATCH /memory/:id update\n DELETE /memory/:id soft delete\n DELETE /memory bulk soft delete (body: {\"ids\":[...]})\n GET /memory/feed.atom Atom 1.0 feed\n GET /memory/feed.json JSON Feed 1.1 (mesh substrate)\n\nTasks\n POST /tasks create\n GET /tasks list\n GET /tasks/:id get task\n PATCH /tasks/:id update\n POST /tasks/:id/claim claim\n POST /tasks/:id/unclaim unclaim\n POST /tasks/:id/complete complete\n POST /tasks/:id/fail fail\n GET /tasks/:id/comments list comments\n POST /tasks/:id/comments add comment\n\nMessages\n POST /messages send\n GET /messages list all sent/received (?read=true|false&limit=)\n GET /messages/inbox unread inbox\n POST /messages/inbox/read-all mark all read\n GET /messages/:id get message by ID\n POST /messages/:id/read mark one read\n\nProjects\n POST /projects create and join\n GET /projects list\n GET /projects/mine your projects\n POST /projects/:id/join join\n DELETE /projects/:id/leave leave\n\nFeeds\n GET /feeds list subscriptions\n POST /feeds subscribe\n PATCH /feeds/:id update name/tags/interval\n DELETE /feeds/:id unsubscribe\n\nMesh\n GET /mesh/peers list linked peers\n POST /mesh/peers link a peer\n DELETE /mesh/peers/:id unlink\n POST /mesh/peers/:id/sync sync now\n GET /mesh/discovered LAN peers via mDNS\n POST /mesh/link-discovered link a discovered peer\n POST /mesh/handshake mutual handshake (unauthenticated, RFC 1918 only)\n GET /mesh/tokens list mesh tokens\n POST /mesh/tokens create token\n PATCH /mesh/tokens/:id update token\n DELETE /mesh/tokens/:id revoke token\n\nAgents\n POST /agents/register register\n PATCH /agents/me rename self\n PATCH /agents/:id rename any (owner)\n DELETE /agents/:id delete (owner)\n GET /agents list with presence (api_key shown to owner only)\n GET /onboard onboarding script\n\nLogs\n POST /logs write log entry (agent+)\n GET /logs list entries (owner)\n\nOAuth (for MCP clients that require it)\n GET /.well-known/oauth-authorization-server\n POST /oauth/register dynamic client registration\n GET /oauth/authorize authorization code + PKCE\n POST /oauth/token token endpoint\n\nOther\n POST /events emit event\n GET /events/stream SSE stream\n POST /sessions/handoff save handoff\n GET /sessions/handoff load handoff + memory delta (your own)\n```\n\n---\n\n## Configuration\n\n| Variable | Default | Description |\n|----------|---------|-------------|\n| `AGENT_KEYS` | | `agent-id:api-key` pairs, comma-separated. Optional `:proj1;proj2` suffix scopes an agent to projects. |\n| `REGISTRATION_KEY` | | Required to register agents (leave blank to disable open registration) |\n| `DB_PATH` | `artel.db` | SQLite path |\n| `PUBLIC_URL` | | Base URL for onboard script and OAuth metadata |\n| `UI_PASSWORD` | | Web UI password |\n| `UI_AGENT_ID` | `artel-ui` | Dashboard agent, auto-created on startup |\n| `UI_DEFAULT_THEME` | `gruvbox` | Default UI theme for new sessions. Options: `gruvbox`, `tokyo-night`, `nord`, `dracula`, `kanagawa`, `rose-pine`, `everforest`, `monokai`, `cobalt`, `solarized`, `hacker`, `mellow`, `volcano`, `ayu`, `flexoki`, `oxocarbon` |\n| `ARCHIVIST_PROVIDER` | `anthropic` | LLM provider: `anthropic` or `openai` |\n| `ARCHIVIST_MODEL` | | Defaults to `claude-sonnet-4-6` / `gpt-4o` |\n| `ARCHIVIST_API_KEY` | | Falls back to `ANTHROPIC_API_KEY` for Anthropic |\n| `ARCHIVIST_BASE_URL` | | OpenAI-compatible base URL (Ollama, Mistral, etc.) |\n| `SYNTHESIS_INTERVAL` | `3600` | Seconds between archivist synthesis passes |\n| `DECAY_RATE` | `0.9` | Confidence multiplier per decay cycle |\n| `DECAY_WINDOW_DAYS` | `7` | Days before decay applies to unmodified entries |\n\n---\n\n## Development\n\n```bash\nuv sync --dev\nuv run pytest tests/ -v\n```\n\n---\n\n## License\n\nMIT. See [LICENSE.md](https://github.com/NicolasPrimeau/artel/blob/master/LICENSE.md).\n","changelogMarkdown","videoKey","requiredSecretsJson","[\"X_AGENT_ID\",\"X_API_KEY\"]","publishedAt","priceCents","trialDays",0,"categoryDisplayName","Developer Tools","upstream",{"_98":99,"_100":101,"_102":103,"_104":105,"_106":107},"requiredSecrets",[96,97],"tags",[],"assets",[],"stats",{"_94":81,"_95":81},"ratingAvg","reviewCount","X_AGENT_ID","X_API_KEY","topics",[108,109,110,111,112,113,114,115,116,117,118],"license","MIT","language","Python","registryName","io.github.NicolasPrimeau/artel","homepage","https://glama.ai/mcp/servers/NicolasPrimeau/artel","agent-coordination","agent-memory","ai-agents","fastapi","llm","mcp","mcp-server","multi-agent","persistent-memory","python","self-hosted","ENV",{"_121":122,"_123":124,"_125":126},"ENVIRONMENT","production","TURNSTILE_SITE_KEY","0x4AAAAAAB8-rEI2gNDf_QhU","SENTRY_DSN","https://45fd1229f78089d5c3ee78be128030e1@o4510583261757440.ingest.de.sentry.io/4510584253448272"]