Not a canned response and not a fixture. Each check makes three live calls: RYO's analyze_token for market facts, DexScreener for the depth of every pool trading the token's canonical contract, and Jupiter for an executable quote to sell your exact size. Then it runs the same pure assess() the CLI uses. When a quote exists, the verdict rests on it and the model's own figure is shown beside it, error and all. If any of the three is down, the answer comes back from the others and says which one failed. The response is RYO's own public envelope, so RYO's clients can read it unchanged.
Pick a token and a size, then run a live check. Usually 2 to 8 seconds: three real network calls.
Every figure above is computed from what this call actually returned. Nothing is filled in when a source doesn't supply it. Solana tokens get an executable quote; ETH and BTC are measured through their wrapped forms on Ethereum, where no keyless quote venue exists, so their cost is modelled and labelled as such. Also callable the way RYO's own tools are: POST /api/tools/exit_liquidity/call, or as an MCP server at /api/mcp. See the exit_liquidity docs.