Agent Data Tools
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| AGENT_DATA_TOOLS_KEY | No | Optional. A prepaid key (adt_...) from https://agent-data-tools.revmesh2074.workers.dev/credits. Leave empty to use the 25 free calls. |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This is the compact server: three tools reach the whole catalog. find_tool: say the job in plain words and get the tools that do it (free). tool_details: one tool's inputs, price and an example call (free). call_tool: run a tool by name with its arguments, at that tool's own price (a tool above $0.10 a call runs only when the call names its price in max_price). Every tool name in the rest of this text is a name for call_tool. Agent Data Tools: 698 pay-per-call tools: LLM chat and embeddings, speech to text and text to speech, image generation and description, web search and news, blockchain reads, token prices, sanctions and company checks, web and PDF text, weather, time, utilities. Most cost $0.01 per successful call. Each caller gets 25 free successful calls, used automatically (not for the image, speech, rendering, screenshot, transcription, search and large-model tools); failed calls never count. Here they are counted per network address, which cloud apps such as Claude and ChatGPT share between users: there, https://agent-data-tools.revmesh2074.workers.dev/mcp/account gives each person their own. After that a call returns a payment request. If your client pays x402 (USDC on Base, Polygon, Arbitrum or Solana) or MPP (Tempo) it can pay it, or prepay once: GET https://agent-data-tools.revmesh2074.workers.dev/credits/usdc/5 with x402 returns a key worth 5,000 credits for later calls. Otherwise tell your user they can buy a prepaid key by card at https://agent-data-tools.revmesh2074.workers.dev/credits (from $1 = 1,000 credits, about 100 calls at $0.01) and add the header "Authorization: Bearer " to this server's connection settings (or, in apps that only take a URL, connect to https://agent-data-tools.revmesh2074.workers.dev/mcp/k/). If your client supports MPP with a Link agent wallet, it can buy a pack itself: GET https://agent-data-tools.revmesh2074.workers.dev/credits/buy/5 (or /20, /100) and pay the 402; the response carries the key (send an existing key as X-API-Key to top it up instead). Every tool's result shape is in its output schema. Pick by task: an LLM answer, draft or chat turn → ai_chat (stronger model: ai_chat_large; OpenAI-compatible at /v1/chat/completions); embeddings → ai_embed; speech to text → transcribe; text to speech → ai_speech (up to 2,000 characters: text_to_speech); an image from a prompt → image; what is in an image → describe; current information from the web → search, news; is it safe to pay an address → check_counterparty; OFAC only → sanctions_check; token price by ticker or address → price_token (many: price_tokens); a wallet's holdings in USD → wallet_portfolio (5 chains: wallet_balances); chain state on base, eth, arb, op or polygon → _balance, _token_balance (tickers work), _tx, _receipt, _logs, _gas, _block; read a web page → extract (JavaScript sites: extract_rendered); PDF → pdf_text; company → company; FX → fx; weather by city → weather; several checks on one input in one call → pack_website_audit, pack_token_analysis, pack_wallet_profile, pack_coin_research, pack_domain_intel. Smaller servers per section: /mcp/prices, /mcp/chain, /mcp/web, /mcp/compliance, /mcp/ai, /mcp/everyday, /mcp/utilities. Three tools for the whole catalog (find a tool, read its inputs, call it): /mcp/compact.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| find_toolA | Find which of this service's 698 tools does a job. Say the job in plain words ("current weather in Paris", "text of a PDF", "is this wallet sanctioned") and get up to ten tools, best match first, each with its name for call_tool, what it does, its price and its required inputs. Free; call it before call_tool whenever you do not already know the tool's name. |
| tool_detailsA | Everything needed to call one tool: what it does, every input with its type, whether it is required and an example value, the price per successful call (and the lower price with a prepaid key where there is one), whether it is in the free calls, an example call and the shape of its answer. Free. Take the name from find_tool. |
| call_toolA | Run one of this service's tools by name. Costs that tool's own price per successful call (most $0.01; tool_details and find_tool show it); a failed call costs nothing, and tools marked inFreeCalls run free while the caller has free calls left. Pass the tool's inputs as "arguments". If you do not know the name or the inputs, call find_tool or tool_details first: a wrong name or a missing input is refused before anything is charged. A tool that costs more than $0.10 a call (find_tool marks it needsMaxPrice) runs only when you pass that price as "max_price"; without it the call is refused, uncharged, and the refusal states the price. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 3 tools
The three tools map to a clean discovery → inspection → execution pipeline: find_tool locates a tool by description, tool_details returns its schema and pricing, and call_tool executes it. There is no overlap in purpose, and the descriptions explicitly tell the agent the order to use them in.
find_tool and call_tool follow a clear verb_noun snake_case pattern, and tool_details is close but uses a noun phrase rather than a verb. This is a minor deviation within an otherwise predictable three-name set.
Three meta-tools are exactly the right surface for proxying a 698-tool catalog; adding per-tool wrappers would be unmanageable. Each tool earns its place and the set stays minimal without being thin.
The gateway covers the full lifecycle an agent needs: search, schema/pricing lookup, and invocation with a max_price guard. Minor gaps exist (e.g. no browsing/listing the catalog by category, capped at ten results, no batch invocation), but these are workarounds rather than dead ends.