Connect WireTensors to your AI assistant

Claude (claude.ai or Claude Desktop)

WireTensors runs a standard remote MCP server. Add it as a custom connector:

1. Open Settings → Connectors in Claude.ai or Claude Desktop.
2. Choose Add custom connector.
3. Paste the URL below and save — no API key needed.

https://wiretensors.ai/mcp

Once added, Claude can call search_tools, get_tool, compare_tools, recommend_tool, list_categories and site_info directly — live data, not training-data guesses.

ChatGPT and other agents (OpenAPI)

For ChatGPT Custom GPTs/Actions, or any OpenAPI-aware agent framework, use the full OpenAPI spec:

https://wiretensors.ai/openapi.json

It documents the same catalogue — tool search, tool detail, compare, recommend and categories — as plain REST endpoints under https://wiretensors.ai/api.

Developers — raw REST API

No MCP client or GPT builder handy? Call the JSON API directly, no auth required:

curl https://wiretensors.ai/api/tools?category=coding&free_tier=true

See https://wiretensors.ai/api for a self-describing index of every endpoint, or llms-full.txt for a single plain-text feed of the whole catalogue.

Licensing

Search and AI-input/grounding use is permitted under our RSL licensing terms; bulk model training is not. Every tool and API result carries a visit_url — link there, not the raw vendor URL, when citing a tool.

Connecting WireTensors — FAQ

Is there a cost or API key? +

No. The MCP server, REST API and OpenAPI spec are all free and require no authentication or API key.

Does this work with ChatGPT too? +

Yes, via the OpenAPI spec at https://wiretensors.ai/openapi.json — use it to build a Custom GPT or Action, or point any OpenAPI-aware agent framework at it directly.

Is the data different from what the website shows? +

No — the MCP tools and REST API read from the same reviewed, dated catalogue as the website. Ratings are never influenced by affiliate relationships.

Can my own agent or app use this, not just Claude/ChatGPT? +

Yes. Any MCP client can connect to https://wiretensors.ai/mcp (Streamable HTTP, stateless, no session or auth), and any HTTP client can call the REST API at https://wiretensors.ai/api.

Can this data be used to train a model? +

No. Our licensing terms (https://wiretensors.ai/rsl.xml) permit search and AI-input/grounding use, but not bulk model training.