canli-mcp / The desk, built into canli-mcp
One tool over hundreds of companies
Run one read-only tool over up to 500 companies at once and get one ranked table of the fields you choose, with medians and any failures, instead of hundreds of separate answers. Long runs come back in parts with a resume ref.
The tool is swarm in canli-mcp 0.22.0, free and MIT-licensed, with no account and no API key. It is part of canli-mcp itself, so it is there as soon as canli-mcp is added.
Arguments
| Argument | Required | Type |
|---|---|---|
tool | yes | string |
companies | no | string[] |
over | no | {tool,arguments,column} |
arguments | no | object |
argument | no | string |
fields | no | string[] |
sort_by | no | string |
order | no | desc|asc |
limit | no | integer |
concurrency | no | integer |
seconds | no | integer |
resume | no | string |
How a model reaches it
Ask in plain words. The model finds swarm with canli-mcp's find_tool, reads its arguments with describe_tool if it needs to, and runs it with run_tool. The call has this shape, with your values in place of the angle brackets:
{
"name": "swarm",
"arguments": {
"tool": "<string>"
}
}Words that find it
Besides its name, title and description, find_tool matches a question to swarm on these words: many, companies, whole, index, universe, bulk, batch, rank, all, sp500, fan, out, parallel, sub, agents, workers, map, scan, every.
More
Nothing here is investment advice.