Features

More than chat, on every uncensored model.

Tools that run inside the request, documents any model can read, a router over the uncensored tiers, a cache for long chats, session traces and the endpoints around them. One key, one prepaid balance.

Get a key

Tool fees, live

Web search
$0.01 per search
Web fetch
$0.001 per page fetched
Code sandbox
$0.05 per container-hour, 1-minute minimum
Remote MCP
No tool fee
Models with tools
41 listed builds

On the catalog now

Uncensored chat
170 models
Auto router
16 models in 3 tiers
File reading
23 formats
free on every model
Embeddings
8 models
Speech to text
1 models

Also on every key

  • Hosted agents: webhook skills, memory and cron on any listed model.
  • Workflows: a JSON graph of model calls run by one request.
  • Daily model check: each model tested against the build it is named after.
  • Signed receipts: every request and monthly statement, verifiable.
  • Playground: any listed model in the browser.
  • Anthropic Messages format: every model answers /v1/messages, so Claude Code and the Anthropic SDK run an uncensored build.
  • MCP server: api.pentestllm.ai/mcp turns every model on the key into tools for Claude Code, Cursor and any MCP client.
  • Sign in with PentestLLM: your app’s users approve a key for it on their own balance, with the prepaid budget your app asks for (OAuth with PKCE).

Request shapes are in the console docs and the OpenAPI document.

Help