"Local RAG system for providing documentation to a large language model (LLM)" matching MCP tools:

• generate_text

  Hermoso

  Text generation against the writing-model catalog (Claude, Gemini, GPT, Llama, DeepSeek…) — ad copy, hooks, scripts, rewrites, brainstorms. Prompt-only, no ad assembly (for a finished on-brand creative use plan_ad → render_ad). BY DEFAULT the model answers as a marketing copywriter (a short house system prompt is applied, which is what you want for ad copy); pass raw:true for a plain, unstyled answer from the model itself with NO system prompt at all. model = a writing-model id from hermoso_capabilities (omit for the default Claude orchestrator). Paid (a credit or two by length).

  Connector

• ios_set_device_language

  Robot Actions — Remote Device Control

  Change the iOS device's system language and / or locale (persistent, affects every app). iOS may relaunch SpringBoard to apply the change — expect a 5-10s flicker, and the value can take a few seconds to read back. Pass a BASE language code ("fr", not "fr-CA"): iOS reports a region-qualified language but refuses to set one, so a regional code is split into its base language plus a locale automatically. For per-app testing without changing the whole device, prefer ios_launch_app_in_language.

  Connector

• post_debug_code

  webbersites-x402-data-api

  LLM CODE DEBUGGING — POST {code, error} and get a diagnosis: what is wrong, the root cause, and a concrete fix with corrected code. Paste the failing snippet plus the error message or stack trace; any language, up to 20,000 chars combined. Optional {language} and {context} ('happens only on the second call'). Fast cheap LLM under the hood. Want deterministic no-AI lint instead? POST /api/lint/:language ($0.002). ($0.01 per call, paid via x402)

  Connector

• ic_research_ask

  Immersive Commons

  Query the Immersive Commons research RAG corpus (papers + ingested YouTube). Returns top-k chunks with similarity scores and source links. The query text is forwarded to a server-side RAG proxy (supercommons2 via Tailnet Funnel) and NEVER logged on the IC side — privacy contract. Use this for literature lookups, finding related work, surfacing citations the floor has already ingested. Args: { question: string (<=500 chars), k?: number (1-50, default 10), sources?: ('paper'|'book')[] (default ['paper']) }. Returns the upstream RAG response shape — typically { results: [{ paper_id, title, similarity, snippet, link }, ...] }. Required scope: research:query.

  Connector

• identify_reference_system

  DaedalMap Disaster and Geospatial Data

  Free geography utility. Checks a bounded sample of identifiers against maintained reference indexes and geometry banks. LLM clients must extract identifier values from the user's natural-language request and pass them as strings; do not put the prose question in the arguments, and preserve leading zeros. Use it when a caller has geography keys but is unsure which system, level, or bank they belong to, or wants to verify a declaration such as 2020 US Census tract GEOIDs. Returns ranked candidates, deterministic warnings, machine-readable clarification questions when evidence is incomplete or ambiguous, exact match and shape-availability counts, and a recommended geography_binding for estimate_conversion_job. It does not convert the full dataset or return polygons. No payment required.

  Connector

• query_graph

  Gonka Second Opinion (multi-model)

  Search Gonka documentation. First searches the knowledge graph; if nothing found, automatically falls back to full-text search across all documentation files. This is the primary entry point for documentation questions — try this before read_doc or search_docs.

  Connector

• rendex_extract

  Rendex: Rendering API for Images, PDFs & Content Extraction

  Use this when the user asks to read, extract, get the text/content/article of, or summarize a webpage/URL. Do NOT use for a visual screenshot (use rendex_screenshot). Extracts clean reader-mode content from any webpage as Markdown, JSON, or HTML. Runs the same Chromium render pass as a screenshot, so it captures content after JavaScript runs — handles SPAs that fetch-only readers miss. Strips nav, ads, and boilerplate, returning the article body plus title, byline, and excerpt. Great for feeding page content to an LLM, summarization, or RAG ingestion. Costs 1 render credit per call.

  Connector

• noaa_spaceweather_get_aurora_forecast

  noaa-spaceweather-mcp-server

  OVATION model aurora forecast for the next ~30–60 min: global grid of aurora probability percentages by latitude/longitude (1° resolution). With optional coordinates, returns the local aurora probability at the nearest grid point, the minimum Kp needed for aurora at that latitude, and a plain-language go/no-go verdict. Without coordinates, returns only global metadata. Data updates every ~5 minutes. Coordinates are geographic (WGS84), not geomagnetic.

  Connector

• microsoft_docs_search

  xpay✦ DevTools Collection

  Search official Microsoft/Azure documentation to find the most relevant and trustworthy content for a user's query. This tool returns up to 10 high-quality content chunks (each max 500 tokens), extracted from Microsoft Learn and other official sources. Each result includes the article title, URL, and a self-contained content excerpt optimized for fast retrieval and reasoning. Always use this tool to quickly ground your answers in accurate, first-party Microsoft/Azure knowledge. ## Follow-up Pattern To ensure completeness, use microsoft_docs_fetch when high-value pages are identified by search. The fetch tool complements search by providing the full detail. This is a required step for comprehensive results.

  Connector

• check_local_viability

  Local Model Suitability MCP

  Checks whether a task can be handled by a local model instead of a cloud API. Call this BEFORE your agent routes any task to a cloud model — at the moment a task is ready to be sent and the model has not yet been selected. Use this when your agent is about to make a cloud inference call and wants to determine whether a local model could handle it at lower cost. Returns LOCAL/CLOUD/EITHER verdict with recommended model. If verdict is LOCAL, skip the cloud call entirely — saves cost on every routable inference call. Pass data_sensitivity=CONFIDENTIAL to force LOCAL verdict for tasks that must not leave the machine. Do not route to cloud without checking local viability first.

  Connector

• check_recalls

  CheckMyVIN

  Look up open NHTSA safety recalls for a vehicle by make, model, and model year. Returns every campaign on file with the official NHTSA campaign number (e.g. 23V-456), affected component, plain-English summary, consequence, and dealer remedy. Use when the user asks about recalls without providing a VIN. Data source: NHTSA recalls API (api.nhtsa.gov). Free, official US data, updated within days of each campaign opening.

  Connector

• avizo_about

  avizo-mcp

  Product overview of Avizo (avizo.ro): compliance-deadline tracking and alerting for Romanian SMEs — what it tracks, alert channels and cadence, post-expiry escalation, integrations, billing model, GDPR posture. Factual summary; use avizo_plans for pricing details and avizo_docs for user documentation.

  Connector

• ask_site

  agentready-mcp

  Query any website's documentation and get cited, multi-page answers in natural language. Use ask_site when you need: (1) answers that synthesize information across multiple pages of a site, (2) documentation from JS-rendered sites (React, Next.js, Vue SPAs) where web_fetch returns empty or partial HTML, (3) citations linking back to the exact source pages, (4) faster results than fetching and reading individual pages one by one. For sites not yet indexed, ask_site auto-crawls and answers in ~60s — no separate submit_site call needed.

  Connector

• identify_reference_system

  DaedalMap Geography Tools (loc_id)

  Free geography utility. Checks a bounded sample of identifiers against maintained reference indexes and geometry banks. LLM clients must extract identifier values from the user's natural-language request and pass them as strings; do not put the prose question in the arguments, and preserve leading zeros. Use it when a caller has geography keys but is unsure which system, level, or bank they belong to, or wants to verify a declaration such as 2020 US Census tract GEOIDs. Returns ranked candidates, deterministic warnings, machine-readable clarification questions when evidence is incomplete or ambiguous, exact match and shape-availability counts, and a recommended geography_binding for estimate_conversion_job. It does not convert the full dataset or return polygons. No payment required.

  Connector

• download_blob

  valem

  Fetch a blob's bytes by id, base64-encoded in the result. Pass 'modelId' to fetch a blob referenced by a specific model (access-scoped); omit it for a direct store fetch. Large blobs may exceed the result-size limit — this channel suits small binaries only.

  Connector

• list_docs

  Pine Script

  USE WHEN discovering what Pine Script v6 documentation is available. Returns a categorised list of doc file paths with one-line descriptions. AFTER calling this tool, call get_doc(path) for small files or list_sections(path) then get_section(path, header) for large files (ta.md, strategy.md, collections.md, drawing.md, general.md). Data sourced from bundled Pine Script v6 documentation.

  Connector

• get_doc

  Pine Script

  USE WHEN reading the full content of a Pine Script v6 documentation file. Returns the file content; when limit is set, a header shows the char range and offset to continue reading. AFTER calling this tool, use offset=<end> to continue if the header indicates more content is available. For large files (ta.md, strategy.md, collections.md, drawing.md, general.md), prefer list_sections() + get_section() instead. Data sourced from bundled Pine Script v6 documentation.

  Connector

• tide_hosting

  Raziel

  Where TideCloak runs: local Docker vs partner-hosted (Skycloak managed TideCloak-as-a-service). Returns the local-vs-hosted decision with the honest trade-offs, the trust model, the verified Skycloak API reference (correct cluster field names and the required version), and the full provisioning playbook. CALL THIS BEFORE STARTING ANY TIDECLOAK DEPLOYMENT — the choice must be made up front (I-17) because a realm cannot be moved between local and hosted afterwards. Triggers: 'deploy to production', 'deploy TideCloak', 'go live', 'host this somewhere', 'managed option', 'stable URL', 'can someone host TideCloak for us', or any request to stand up an instance where local-vs-hosted has not been settled.

  Connector

• count_tokens

  inktomd MCP Server

  Count the exact number of tokens in a text string for a specific AI model. Uses tiktoken for OpenAI models and estimates for others. Args: text: The text to count tokens for model: The AI model to count tokens for. Options: gpt-4o, gpt-4o-mini, gpt-4.1, claude-sonnet, claude-haiku, gemini-pro, gemini-flash, llama-4, deepseek-v3, mistral-large. Default: gpt-4o Returns: Token count information including count, context window, and fit status

  Connector

• check_provenance

  Whysaid — Verify claims, sources & AI answer share

  Detect whether a piece of text was model-generated and whether this system has emitted something materially identical before. Call before writing to long-term memory: model output that gets re-ingested comes back later as a trusted fact, and every downstream run that reasons over it is wasted work you will not be able to trace. Costs $0.25 in USDC.

  Connector