Docs Mcp

frumu-ai/tandem

This is a streamable HTTP endpoint that gives Claude direct access to Tandem's runtime documentation, including installation guides, SDK references, workflow examples, and agent setup instructions. If you're integrating Tandem's governed runtime into your AI tooling or trying to understand how its authority layer, approval gates, tenant-aware memory, or MCP connector system works, this server lets Claude pull the relevant docs without you switching tabs. It's particularly useful when you're configuring workflows, setting up permissioned knowledge spaces, or debugging multi-tenant deployments where the runtime needs to enforce scoped execution and audit trails below the model layer.

GitHub stars: 107
tools: 13
transport: HTTP
last updated: Apr 22, 2026

View on GitHub

Install

From registry metadata

$ npx add-mcp https://tandem.ac/mcp --name docs-mcp

Run in your terminal. add-mcp asks which agents to add it to (Claude Code, Codex, Cursor and more). Add -g to install it for every project.

Review the command, arguments and environment values before installing. MCP servers run with your local permissions.

Other ways

Claude Code

$ claude mcp add --transport http docs-mcp https://tandem.ac/mcp

Run in your terminal. Add --scope user to make it available in every project.

Tools

Verified live · 13 tools

Verified live against the running server on Nov 6, 2026.

search_docsSearch the published Tandem docs index for relevant pages.5 params

Search the published Tandem docs index for relevant pages.

Parameters* required

audiencestring

Optional audience filter such as developer, desktop user, or operator.

docs_refstring

Docs channel or immutable ref. Defaults to stable. Use next, refs/<sha>, or releases/<version> only when explicitly needed.

engine_versionstring

Optional Tandem engine version hint. If supplied without docs_ref, the MCP still defaults to stable unless a matching versioned docs ref is configured.

limitinteger

Maximum number of matching docs pages to return.

query*string

The Tandem question, task, or keyword query to search for.

get_docFetch and normalize a published Tandem docs page by path or URL.3 params

Fetch and normalize a published Tandem docs page by path or URL.

Parameters* required

docs_refstring

Docs channel or immutable ref. Defaults to stable.

engine_versionstring

Optional Tandem engine version hint for compatibility checks.

path*string

Docs path like /sdk/python/ or a full Tandem docs URL.

get_start_pathRecommend the best Tandem docs starting path for a desktop user, server admin, or developer.2 params

Recommend the best Tandem docs starting path for a desktop user, server admin, or developer.

Parameters* required

docs_refstring

Docs channel or immutable ref. Defaults to stable.

role*string

The user role to optimize the docs starting point for.one of desktop · server_admin · developer

recommend_next_docsRecommend the next Tandem docs to read after a current page, optionally biased toward a goal.4 params

Recommend the next Tandem docs to read after a current page, optionally biased toward a goal.

Parameters* required

current_path*string

Current Tandem docs path the user has already read.

docs_refstring

Docs channel or immutable ref. Defaults to stable.

goalstring

Optional goal such as install Tandem, use the Python SDK, or build workflows.

limitinteger

Maximum number of next-doc recommendations to return.

answer_how_toAnswer how to install or use Tandem, build workflows, or use the SDKs by synthesizing relevant published docs.5 params

Answer how to install or use Tandem, build workflows, or use the SDKs by synthesizing relevant published docs.

Parameters* required

audiencestring

Optional audience hint such as developer, desktop user, or operator.

docs_refstring

Docs channel or immutable ref. Defaults to stable.

engine_versionstring

Optional Tandem engine version hint. Prefer stable docs for installed engines.

limitinteger

Maximum number of supporting docs recommendations to include.

task*string

The Tandem task or question that needs a docs-grounded answer.

get_tandem_guideReturn a structured Tandem learning guide for install-engine, python-sdk, typescript-sdk, build-workflows, or headless-deploy.2 params

Return a structured Tandem learning guide for install-engine, python-sdk, typescript-sdk, build-workflows, or headless-deploy.

Parameters* required

docs_refstring

Docs channel or immutable ref. Defaults to stable.

topic*string

Named Tandem learning guide to return.one of install-engine · python-sdk · typescript-sdk · build-workflows · headless-deploy

warmup_docs_cacheRefresh the docs index and preload the most important Tandem docs pages into cache.2 params

Refresh the docs index and preload the most important Tandem docs pages into cache.

Parameters* required

docs_refstring

Docs channel or immutable ref. Defaults to stable.

force_refreshboolean

Whether to bypass cache and force fresh upstream fetches first.

get_docs_cache_statusInspect the current docs cache state for index sources and warmed pages.

Inspect the current docs cache state for index sources and warmed pages.

No parameters. Call it with no arguments.

refresh_docs_indexForce-refresh docs indexes and report whether data came from the network, cache, or stale fallback.2 params

Force-refresh docs indexes and report whether data came from the network, cache, or stale fallback.

Parameters* required

docs_refstring

Docs channel or immutable ref. Defaults to stable.

force_refreshboolean

Whether to bypass cache and force a network refresh attempt.

refresh_doc_pageForce-refresh a docs page and report whether content came from the network, cache, or stale fallback.3 params

Force-refresh a docs page and report whether content came from the network, cache, or stale fallback.

Parameters* required

docs_refstring

Docs channel or immutable ref. Defaults to stable.

force_refreshboolean

Whether to bypass cache and force a network refresh attempt.

path*string

Docs path like /start-here/ or /sdk/python/ to refresh.

invalidate_docs_cacheInvalidate docs cache entries for index data, all pages, or a specific page path.3 params

Invalidate docs cache entries for index data, all pages, or a specific page path.

Parameters* required

docs_refstring

Optional docs ref to scope page/index cache invalidation.

pathstring

Required when target is page; identifies the specific docs path to invalidate.

targetstring

Which cache area to invalidate.one of all · index · pages · page

compare_docs_index_refreshForce-refresh the docs indexes and return before/after hashes so content changes can be proven.1 params

Force-refresh the docs indexes and return before/after hashes so content changes can be proven.

Parameters* required

docs_refstring

Docs channel or immutable ref. Defaults to stable.

compare_doc_page_refreshForce-refresh a docs page and return before/after hashes so page content changes can be proven.2 params

Force-refresh a docs page and return before/after hashes so page content changes can be proven.

Parameters* required

docs_refstring

Docs channel or immutable ref. Defaults to stable.

path*string

Docs path like /start-here/ or /sdk/python/ to compare before and after refresh.

Readme

Tandem Logo

View the full README on GitHub

Details

Repository: github.com/frumu-ai/tandem
Website: tandem.ac/docs-mcp
Category: AI & LLM Tools
Transport: HTTP
Registry: MCP registry, active
Resources: 50
Prompts: 3
Tools verified: Nov 6, 2026

Related AI & LLM Tools MCP servers

View all

Pensyvemajor7apps/pensyveUniversal memory runtime for AI agents — episodic, semantic, and procedural memory.99stars

emem, the verifiable memory protocol for the physical worldvortx-ai/ememShared, verifiable memory for machines and AI. Encode where data lives; decode with any AI.89stars

Sverklosverklo/sverkloLocal-first MCP code intelligence: 37 tools — hybrid search, blast-radius, diff review, memory.71stars

Token Enhancerxelektron/token-enhancerCuts AI web-fetching costs up to 99.9% by stripping page junk before it reaches your LLM.66stars

Threadline Mcpsagemindai/instarAgent-to-agent messaging relay. Discover agents, send messages, and receive replies. Zero-config.0stars

Three.js DevTools MCPdmitriygolub/threejs-devtools-mcpInspect and edit Three.js scenes, materials, shaders, lights in real time from any AI agent64stars