Skip to main content
Glama

Tanilo MCP server

An MCP server that connects Claude Desktop, Cursor, Windsurf or any other MCP client to the Tanilo API at https://api.tanilo.io.

Tanilo checks the premise before an AI agent acts. This server exposes two checks, the contract price check and the spend cap check, and two status documents. The routes it calls are live in free beta, rate-limited, no key needed. The server holds no secrets and makes no payments.

Tanilo was AgentOracle until September 2026. Versions up to 2.1.2 were published as agentoracle-mcp; that package is no longer maintained.

This README describes version 3.1.0. The outputs below were captured by running the tools against the live API: on 2026-10-08 for the contract price check and on 2026-10-13 for the spend cap check.


Install

npx tanilo-mcp

Claude Desktop

Open your Claude Desktop config file:

Mac: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json

Add this and restart Claude Desktop:

{
  "mcpServers": {
    "tanilo": {
      "command": "npx",
      "args": ["-y", "tanilo-mcp"]
    }
  }
}

Cursor (Settings → MCP → Add Server) and Windsurf (Settings → MCP Servers → Add) take the same tanilo block.


Related MCP server: MCP Perplexity Pro

Tools

check-contract-price

Does this offer match the terms you supplied? Before an agent accepts an offer, contract_price_match compares the offer's unit price with the applicable term among the terms you supply, and signs the result when signing succeeds.

No model is involved: the result is arithmetic and exact comparison, and anyone holding the same input can recompute it. You supply the terms; Tanilo does not hold or look up your contracts. Three results: verified, contradicted, or indeterminate with a reason. A verified result says the price matches the supplied term. Whether to pay is your policy's decision.

Input: offer (unit_price, currency, unit, and optionally quantity, seller, sku, offered_at) and terms (a list; each has agreement_id, version, currency, unit, and either price_per_unit or tiers, plus optional seller, product_scope, effective_from, effective_to, supersedes). Prices and quantities are decimal strings such as "2.00". The full input rules, tiers and amendments are at https://tanilo.io/docs/contract-price-check.

The tool returns the check result and the signed receipt. Captured live on 2026-10-08 for the docs example (an offer of USD 2.00 per call against a supplied term of USD 1.00 per call), trimmed:

{
  "verdict": "halt",
  "state": "contradicted",
  "outcome": "fail",
  "reason": "price_mismatch",
  "recommendation": "offer price differs from the applicable term among the supplied terms",
  "evidence": {
    "rule_id": "contract-price-match/v0.1",
    "terms_source": "caller_supplied",
    "expected": { "price_per_unit": "1.00", "currency": "USD", "unit": "call" },
    "offered": { "unit_price": "2.00", "currency": "USD", "unit": "call", "quantity": "1" },
    "difference_per_unit": "1.00",
    "offer_is": "above_term_price"
  },
  "check_mode": "deterministic",
  "canonical_sha256": "sha256-…",
  "kid": "tanilo-2026-10-ed25519-7d885da9",
  "jwks_url": "https://tanilo.io/.well-known/jwks.json",
  "receipt": { "payload": "…", "signatures": [ … ] }
}

With "terms": [] the same offer gives state indeterminate, reason no_applicable_term, verdict halt.

A valid signature shows which key signed these bytes and that they have not changed since. It does not show that the supplied terms were genuine. Check the signature offline with tanilo-receipt-verify against the key set at https://tanilo.io/.well-known/jwks.json:

pip install tanilo-receipt-verify
python3 -c "import json; from tanilo_receipt_verify import verify; r = verify(json.load(open('receipt.json')), jwks_by_issuer={'https://tanilo.io/.well-known/jwks.json': json.load(open('jwks.json'))}); print(r.status, r.canonical_sha256)"

where receipt.json holds the tool's receipt object and jwks.json the key set.

check-spend-cap

Does this offer fit within the cap you supplied? Before an agent commits to a purchase, spend_cap multiplies the offer's unit price by its quantity, exactly, and compares the total with the remaining cap you supply. It signs the result when signing succeeds.

You supply the cap and already_spent. This check does not retrieve budget balances, maintain a running balance or reserve funds. Three results: verified, contradicted, or indeterminate with a reason. A verified result says the offer's total is within the remaining cap. Whether to pay is your policy's decision.

Input: offer (unit_price, currency, and optionally quantity, offered_at, seller, sku) and cap (amount, currency, scope of per_purchase or per_period, and for per_period already_spent and period with from and to; optionally cap_id). Optional claim_hash and agent_id as for check-contract-price. The formats are on the docs page: https://tanilo.io/docs/spend-cap-check.

The tool returns the check result and the signed receipt. Captured live on 2026-10-13 for the docs example (three calls at USD 12.50 against a USD 100.00 cap of which USD 70.00 is already spent), trimmed:

{
  "verdict": "halt",
  "state": "contradicted",
  "outcome": "fail",
  "reason": "cap_exceeded",
  "recommendation": "offer total exceeds the remaining cap the caller supplied",
  "offer_total": "37.50",
  "remaining": "30.00",
  "excess": "7.50",
  "cap_already_exhausted": false,
  "evidence": {
    "rule_id": "spend-cap/v0.1",
    "cap_source": "caller_supplied",
    "offer_total": "37.50",
    "remaining": "30.00",
    "excess": "7.50",
    "cap_already_exhausted": false,
    "offer_sha256": "sha256-…",
    "cap_sha256": "sha256-…",
    "steps": [ … ]
  },
  "check_mode": "deterministic",
  "canonical_sha256": "sha256-…",
  "kid": "tanilo-2026-10-ed25519-7d885da9",
  "jwks_url": "https://tanilo.io/.well-known/jwks.json",
  "receipt": { "payload": "…", "signatures": [ … ] }
}

With "already_spent": "62.50" the same offer is exactly at the cap: state verified, remaining 37.50, excess null, verdict act.

Verify the receipt's signature against a key set you have authenticated as Tanilo's, as for check-contract-price. A valid signature establishes that the signed content verifies under the selected key and has not been altered. It does not establish that the supplied cap is authentic, approved or current, or that the purchase is authorized.

check-health

Returns the API's status document (GET /health): service status, the live routes and their prices, the retired routes, and rate limits.

get-manifest

Returns the API's discovery document (GET /.well-known/x402-manifest.json). It lists no payable resources: the paid research routes were retired on 2026-10-04.


What this server does not do

  • It has no tool for Tanilo's claim check (POST /evaluate) yet. The route is free during the beta and can be called directly; see https://tanilo.io/pricing.

  • It does not pay for anything, hold a wallet, or read any key. The x402 payment code of earlier versions is gone.

  • It does not verify signatures itself; use tanilo-receipt-verify as above.

Development

npm install
npm run build
npm test          # starts the built server over stdio and calls the live API (free beta route)

Set TANILO_PYTHON to a Python that has tanilo-receipt-verify installed and the test also verifies the returned receipt's signature offline. Set TANILO_API_BASE to point the server at another host (tests only).


Contact: joe@tanilo.io

MIT licensed. TK Collective LLC.

Available Tools

4 tools
check-healthA

Check if the AgentOracle API is online and get service status including uptime, model info, and pricing.

Inputs: no named parameters.

deep-researchA

Perform comprehensive deep research on any topic using AgentOracle Sonar Pro. Returns expert-level analysis with 10-15 detailed facts, in-depth analysis paragraph, cited sources, and confidence score. Costs $0.10 USDC per query via x402 on Base mainnet. Use for complex topics requiring thorough analysis.

Inputs: query (required).

get-manifestA

Get the AgentOracle x402 payment manifest. Returns payment requirements, supported networks, pricing, and endpoint details for programmatic integration.

Inputs: no named parameters.

researchA

Perform real-time research on any topic using AgentOracle. Returns structured results with summary, key facts, cited sources, and confidence score. Costs $0.02 USDC per query via x402 on Base mainnet.

Inputs: query (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    A4.2/5.0

    Scored across 4 tools

    Disambiguation5/5

    Each tool has a clearly distinct purpose: check-health verifies API status, deep-research provides comprehensive analysis, get-manifest retrieves payment details, and research offers basic real-time research. There is no overlap in functionality, and the descriptions clearly differentiate their scopes and use cases.

    Naming Consistency5/5

    All tool names follow a consistent kebab-case pattern with clear verb-noun combinations: check-health, deep-research, get-manifest, and research. The naming is predictable and readable, with no deviations in style or convention.

    Tool Count5/5

    With 4 tools, the set is well-scoped for the AgentOracle MCP server's purpose, covering API health checks, payment integration, and two tiers of research functionality. Each tool earns its place without redundancy, making the count appropriate for the domain.

    Completeness4/5

    The tool surface covers core functionalities: API status, payment details, and research operations with basic and deep tiers. A minor gap exists in lacking explicit CRUD operations for managing research queries or payment settings, but agents can work around this given the server's focus on querying and integration.

    Maintenance

    ActivitySlowing
    ResponsivenessNo recent community issues

    Related MCP Connectors

    • Web search API and MCP server for developers building AI agents.

      https://searchpipe.tech/mcp/
    • Hosted MCP server for live public-data APIs and Skills for AI agents.

      https://mcp.unifapi.com
    • One MCP server for all web data. Scrape, search, crawl, and AI research in one install.

      https://mcp.novada.com/mcp
    • Your agent needs the open web — searched by more than one engine, and read as clean markdown rather than raw HTML. **What you can ask for** • "Search this question with two providers and tell me where they disagree." • "Scrape these 40 URLs into markdown, in one batch." • "Crawl this documentation site and give me every page." • "Do deep research on this topic and cite the sources." • "Find the academic papers behind this claim." **How to use it** Point any MCP client at https://mcp.aisa.one/search/mcp and sign in with OAuth — there is no key to create or paste. 30 tools across several independent providers: Tavily and Exa search, answers, contents and agent runs; Firecrawl scrape, batch scrape, crawl, map and search; Perplexity Sonar, Sonar Pro, reasoning and deep research; Oxylabs AI search and LLM jobs; OpenAI and Anthropic web search; and scholarly search. **Why this rather than the source** Several independent indexes behind one account, because one engine's blind spot is not visible from inside it. **It is also a door to the rest** The same login reaches 26 sources and 580+ operations. Find the page here, then ask the same agent who links to it or how much traffic it gets — without adding a second server. **What it costs** Finding and inspecting an operation is free. Running one is billed per call at API prices, with no seat and no monthly minimum, and every call takes max_price_usd so an agent cannot overspend by accident. **Where else it reaches** https://mcp.aisa.one/seo-serp/mcp for the Google results page itself, https://mcp.aisa.one/seo-serp-other-engines/mcp for Bing, Baidu and Naver.

    Related MCP Servers

    • A
      license
      A
      quality
      B
      maintenance
      A comprehensive MCP server that provides intelligent access to Perplexity AI's search and reasoning models with automatic model selection, conversation management, and project-aware storage. Supports real-time search, deep research, chat sessions, and async operations for complex queries.
      11
      22 PyPI
      3
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      An MCP server that enables AI agents to perform search-augmented queries and deep multi-source research using the Perplexity API.
      26
      Apache 2.0