Andish
Knowledge-graph memory for your team's documents.
Connector by andish.ruOfficial · Domain verified by the MCP Registry
Install
Works with Claude Code, Codex, VS Code, Cursor, Windsurf, Claude Desktop. Read the repository before you give any tool access to your data.
npx agentmarket install connector/andish-ru-andish-mcp
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this server can do and what it asks for, from its tool list and install config. Checked Oct 25, 2026.
- Tools
- 8 read · 1 write · 0 delete
- Secrets
- 1 (API keys, tokens or passwords)
- Install scripts
- None found
- Runs as
- Python package
Tools by what they do 9
Read from the source code; the running server may list different tools.
Read 8 · Look things up without changing anything.
andish_contextandish_get_documentandish_get_neighborsandish_healthandish_list_documentsandish_queryandish_search_entitiesandish_workspace_usage
Write 1 · Create, change, send or run things.
andish_upload_document
Secrets and settings it asks for
Secrets: values you paste in that let it act as you.
ANDISH_API_KEY
Other settings (environment variables or headers).
ANDISH_BASE_URL
Files it touches
- Any path the agent passes to its tools
Tools (9)
Read from the source code; the running server may differ. Descriptions are the server’s own words; hidden characters are shown as [U+…].
andish_contextFetch raw context from the team's Andish knowledge graph and write the answer yourself. Use this tool when you want to compose the answer, not have it composed for you. It returns the source material — entities, relations, and document fragments with their sources — but no ready-made answer. You spend fewer tokens and reply faster than with andish_query, because Andish does not synthesise a second time. Choosing between the two tools: - andish_context (this tool): you get raw context and write the answer. Cheaper, faster. - andish_query: you get a ready, synthesised answer with sources. …
andish_get_documentFetch metadata and a content excerpt for a specific Andish document. Use this tool when you need details about a particular document — its processing status, upload time, metadata, or a preview of its content. Args: document_id: UUID of the document (e.g. from andish_list_documents). Returns: Markdown-formatted document metadata and content excerpt. Returns a user-friendly error on 404 or API errors.
andish_get_neighborsReturn entities directly (hop=1) or transitively (hop=2) connected to entity_id. Use this tool to explore relationships of a known entity in the knowledge graph. hop=1 returns direct neighbors; hop=2 includes neighbors-of-neighbors. Args: entity_id: Stable entity identifier returned by andish_search_entities. hop: Traversal depth: 1 (direct neighbors) or 2 (2-hop). Default: 1. limit: Maximum number of neighbor entities to return (1-200). Default: 50. Returns: Entity label with a list of neighbors and edge types, or "Entity has no neighbors." when the entity exi…
andish_healthCheck the health of the Andish MCP server and API connectivity. Use this tool to verify the MCP connection is live, the API key is valid, and the Andish backend is reachable. Designed for the Phase 2 onboarding wizard success state — call this tool first to confirm everything is working before running any other Andish tools. No parameters required. Returns: JSON health report with status ("ok" | "degraded" | "down"), version, API reachability, auth validity, LLM provider, database connectivity, server uptime in seconds, and a UTC timestamp. Status "ok" means all systems are…
andish_list_documentsList documents uploaded to the Andish workspace. Returns a markdown table of documents with their ID, name/title, status, and upload date. Only the first page is returned (Sprint 2). To see more documents increase the limit parameter (max 100). Feature 03-bis: use ``filter`` to focus on documents that may need review or cleanup. Args: limit: Maximum number of documents to return (1-100, default 50). filter: Optional activity filter. Allowed values: - ``"never_queried"`` — documents never cited by any query. - ``"quiet_90d"`` — not cited in 90+ days (includes …
andish_querySearch the Andish workspace knowledge graph and synthesise an answer. Use this tool when you need to answer questions based on the user's workspace (documents previously uploaded to Andish). Returns a natural-language answer with source document references. Args: question: Natural-language question in Russian or English (3-2000 chars). mode: Retrieval mode. "mix" (default, recommended) combines knowledge graph + vector search. "local", "global", "hybrid", "naive" are experimental and may return empty answers for sparse graphs. Returns: Synthesised answer text…
andish_search_entitiesSearch entities (people, organizations, concepts) by label substring in workspace. Use this tool when you need to find specific entities in the knowledge graph by name fragment. Returns a list of matching entity identifiers and their types. Args: query: Substring to search in entity labels (case-insensitive, 1-200 chars). limit: Maximum number of results to return (1-100, default: 20). Returns: Formatted list of matching entities or "No entities matched." when empty. Returns a user-friendly error message on API errors (not raw HTTP codes).
andish_upload_documentUpload a local file to the Andish workspace for indexing. Reads the file at ``file_path`` as UTF-8 text and enqueues it for knowledge graph extraction. Use ``andish_get_document`` to poll the returned ``document_id`` until status becomes "indexed". Typical file types: Markdown (.md), plain text (.txt), any UTF-8 file. Maximum size: 1 MB. Content above 15 000 words is split automatically. Args: file_path: Absolute or relative path to the local file. title: Optional display title (overrides filename). document_date: Optional validity date label, e.g. "2024-Q3" or …
andish_workspace_usageShow current Andish workspace usage: documents, cost, and daily queries. No parameters needed — the workspace is determined from the API key. Use this tool to answer questions like "how many documents do I have?" or "how much have I spent today?". Also shows workspace cleanup hints (documents never queried or quiet for 90+ days). Returns: Human-readable summary of workspace usage metrics. Returns a user-friendly error message on API errors (not raw HTTP codes).
Signals
- Weekly downloads
- 26
- Last release
- Aug 16, 2026 (2 months ago)
- Last package publish
- Aug 16, 2026 (2 months ago)
From the package registry.
Details
- Registry name
ru.andish/andish-mcp- Version
- 0.2.0
- Auth
- API key
- Packages
- pypi andish-mcp
- License
- License unknown
- Website
- andish.ru
- First seen
- Oct 24, 2026
- Last checked
- Oct 25, 2026
- Source
- Official MCP Registry
Cite this page
Andish MCP server. Agent Mart. https://agentmart.dev/connectors/andish-ru-andish-mcp (last updated 2026-10-25)