Home › Knowledge & Memory › maxkuminov/obsidian-mcp
maxkuminov/obsidian-mcp
Self-hosted agent memory as a markdown vault: semantic and keyword search, wikilink graph, OAuth
Topics: Knowledge & Memory
Connect
Review any command before running it. Package names and URLs come from the server's own registry entry.
Package (oci)
Package: oci ghcr.io/maxkuminov/obsidian-mcp:v1.4.0
DATABASE_URLrequired secret — PostgreSQL 16 + pgvector (0.8.0+) connection URL for the app's NON-superuser role, e.g. postgresql+asyncpg://obsidian_mcp:<password>@postgres:5432/obsidian_mcp. A superuser role or a CHANGE_ME password is refused at startup.VAULT_PATHrequired — Path of the Obsidian vault inside the container (mount your vault read-write there).SECRET_KEYrequired secret — itsdangerous signer for cookies/sessions. Generate with `openssl rand -hex 32`; the app refuses to start on a placeholder.MCP_HOSTNAMErequired — Public hostname your reverse proxy routes to. Also derives BASE_URL, ALLOWED_ORIGINS and ALLOWED_HOSTS as https://<MCP_HOSTNAME> unless those are set explicitly.EMBEDDING_PROVIDER— Embedding provider: `ollama` (self-hosted, default), `openai` (bring your own key) or `tei` (Hugging Face Text Embeddings Inference, native /embed). Switching after deploy requires `make reset-embeddings`.OLLAMA_URL— Ollama endpoint serving bge-m3 (used when EMBEDDING_PROVIDER=ollama). Must be https or loopback unless EMBEDDING_ALLOW_PLAINTEXT is set.OPENAI_API_KEYsecret — OpenAI (or OpenAI-compatible) API key for text-embedding-3-* (required when EMBEDDING_PROVIDER=openai; the server refuses to start without it).TEI_URL— TEI endpoint (used when EMBEDDING_PROVIDER=tei). Must be https or loopback unless EMBEDDING_ALLOW_PLAINTEXT is set.TEI_EMBEDDING_MODEL— Must equal the model_id the TEI server reports at GET /info, exactly (a local path included); a different value refuses startup. A change needs `make reset-embeddings`.TEI_EMBED_BATCH_SIZE— Inputs per TEI /embed request (1-256). A value above the server's max_client_batch_size refuses startup.
Related servers
modelcontextprotocol/servers
Direct local file system access.
Context7
Up-to-date docs for AI agents
hindsight
Hindsight: long-term memory for AI agents modeled on human memory.
cognee
Memory manager for AI apps and agents using graph and vector stores, with ingestion from many data sources.
Skill_Seekers
Turn docs, GitHub repos, PDFs, videos, Confluence, Notion and Slack/Discord into AI skills and RAG knowledge, exportable to vector…
basic-memory
Local-first knowledge management with bi-directional LLM sync via Markdown files.
Listed in punkpeye/awesome-mcp-servers (MIT)
Data from the Official MCP Registry