ToolPlexerEndpoints
toolplexer run
Execute one inspected capability endpoint.
Run an endpoint after inspecting its schema and price. Execution is billed.
Usage
toolplexer run -p <provider> -e <endpoint> -i <input> \
[--wait [timeout]] [-o <file>] [--json]Flags
| Flag | Type | Description |
|---|---|---|
-p, --provider <provider> | string | Provider slug |
-e, --endpoint <endpoint> | string | Endpoint path |
-i, --input <input> | JSON | Inline JSON, @path/to/file.json, or - for stdin |
-f, --input-file <path> | file | Read JSON from a file; alternative to -i @file |
-w, --wait [timeout] | number | Wait for completion; timeout in seconds, default 300 |
--timeout <seconds> | number | Polling deadline used with --wait (1–3600) |
-o, --output <file> | file | Save the complete response without overwriting |
-j, --json | boolean | Print raw JSON |
The CLI performs a free inspection and required-field check before placing the
paid request. Current adapters accept JSON body input; unsupported nonempty
--query and --path objects are rejected explicitly.
Examples
Start a run and poll separately:
toolplexer run -p mock -e /web-search \
-i '{"query":"cats","maxResults":10}'
toolplexer runs get -r <run-id> -o results.jsonWait for completion:
toolplexer run -p mock -e /web-search \
-i '{"query":"cats","maxResults":10}' --wait -o results.jsonRead complex input from a file:
toolplexer run -p mock -e /web-search -i @params.jsonUse execute when ToolPlexer should select the best-value configured provider
instead of selecting a provider yourself.