# Tako

> Give your agent web search and authoritative datasets: S&P Global, FRED, OECD, SimilarWeb & more.

- name: io.github.TakoData/tako-mcp
- version: 1.4.4
- connection class: R0 (autonomous)
- trust: 94/100
- quality: 91/100 (strong)
- owner: not claimed
- descriptor: https://api.protogrid.dev/v1/servers/io.github.TakoData%2Ftako-mcp
- tools: https://api.protogrid.dev/v1/servers/io.github.TakoData%2Ftako-mcp/tools

**An agent can connect right now, no human step.** Remote endpoint, no authentication, reachable on the last probe.

## Connection (mcpServers)

```json
{
  "mcpServers": {
    "tako-mcp": {
      "type": "http",
      "url": "https://mcp.tako.com/mcp",
      "headers": {
        "Authorization": "Bearer ${TAKO_MCP_TOKEN}"
      }
    }
  }
}
```

Placeholders to fill: `${TAKO_MCP_TOKEN}`.

## Trust

- hygiene: 100
- liveness: 100
- freshness: 100
- provenance: 80
- drivers: +repository +namespace-matches-repo +reachable +uptime-100% +updated-1d-ago
- Derived from observable signals (official registry feed and our own credential-free probes); no code audit performed.

## Quality 91/100 (strong)

### protocol: 75 (weight 25)

- warn `protocol.modern`: newest supported version is 2025-11-25; 2026-07-28 not supported, older versions are deprecated until 2027-07-28
- n/a `protocol.stateless`: only applies to 2026-07-28 servers
- pass `protocol.transport`: streamable HTTP
- n/a `protocol.list_ttl`: only applies to 2026-07-28 servers

### authorization: n/a (weight 15)

- n/a `auth.prm`, `auth.as_metadata`, `auth.cimd`: no remote uses OAuth
- n/a `auth.secret_in_url`: no templated URL

### tool hygiene: 100 (weight 30)

- pass `tools.descriptions`: every tool has a description
- pass `tools.description_length`: descriptions are concise
- pass `tools.schemas`: every tool has an object input schema
- pass `tools.annotations`: 4 of 4 tools declare readOnlyHint or destructiveHint
- pass `tools.directory_hints`: every tool declares readOnlyHint, destructiveHint, idempotentHint and openWorldHint
- pass `tools.token_cost`: about 2,097 tokens to load every tool
- pass `tools.api_dump`: tools are not a one-to-one API dump

### stability: 100 (weight 15)

- pass `stability.changes`: 3 tool changes in 30 days
- pass `stability.rug_pull`: no tool changed its meaning under the same name

### dependencies: n/a (weight 15)

- n/a `deps.known_vulns`, `deps.mcp_sdk_version`, `deps.resolvable`: no npm or PyPI package

- tools: 4, about 2,097 tokens to load them all
- tool set hash: c6b987e5407c7e41bf0c0d947215c21b
- tools last changed: 2026-10-08T02:23:21.219Z
- badge: [![protogrid quality](https://protogrid.dev/badge/io.github.TakoData/tako-mcp.svg)](https://protogrid.dev/servers/io.github.TakoData/tako-mcp)
- Checks run on what our credential-free, read-only probes observe; tools are never called and no code audit is performed.

### Tool changes

- 2026-10-08 `tako_search` changed (description, outputSchema)
  - before: Search Tako's data graph and the live web in one call: many results at once, as structured cards plus web results, with the top card rendered inline as a chart. It finds data; `tako_contents` fetches it. Each card carries a headline value, node ids, and a url — pass the url to `tako_contents` for rows (`exportable: true` cards) or a web result's full page text. When `exportable` is false the rows are locked — read the headline value from the card's `description`. Best for: breadth — fan out several narrow queries in parallel. Each query resolves one metric — for one entity, or a comparison set ("Apple revenue", "Nvidia vs AMD gross margin"); several metrics or topics in one query retrieve poorly. To learn what Tako covers, or a metric's canonical name, run `tako_available_data` first, then search on the canonical name it returns.
  - after: Search Tako's data graph and the live web in one call: many results at once, as structured cards plus web results, with the top card rendered inline as a chart. It finds data; `tako_contents` fetches it. Each card carries a headline value, node ids, and a url — pass the url to `tako_contents` for rows (`exportable: true` cards) or a web result's full page text. When `exportable` is false the rows are locked — read the headline value from the card's `description`. Website-traffic and SEO cards bill a per-unit data minimum above the search price. Best for: breadth — fan out several narrow queries in parallel. Each query resolves one metric — for one entity, or a comparison set ("Apple revenue", "Nvidia vs AMD gross margin"); several metrics or topics in one query retrieve poorly. To learn what Tako covers, or a metric's canonical name, run `tako_available_data` first, then search on the canonical name it returns.
- 2026-10-08 `tako_contents` changed (description)
  - before: Fetch the full content behind a url: a web page's text, or an exportable Tako card's data rows. Batch up to 10 urls in one call — each one is billed and fails on its own. Fetch only cards that `tako_search` marked `exportable: true`. Rows bill per 1,000 delivered, so set `max_rows` when the recent rows are enough. If a page is long, such as a filing or an annual report, set `query` to a question to get back only the passages that answer it. Best for: reading one source in full — a page you need to quote, or the rows behind a card you need to compute over.
  - after: Fetch the full content behind a url: a web page's text, or an exportable Tako card's data rows. Batch up to 10 urls in one call — each one is billed and fails on its own. Fetch only cards that `tako_search` marked `exportable: true`. Rows bill per 1,000 delivered, so set `max_rows` when the recent rows are enough. A website-traffic or SEO card also bills a per-unit data minimum that `max_rows` doesn't bound. If a page is long, such as a filing or an annual report, set `query` to a question to get back only the passages that answer it. Best for: reading one source in full — a page you need to quote, or the rows behind a card you need to compute over.
- 2026-10-07 `tako_contents` changed (description, inputSchema, outputSchema)
  - before: Fetch the full content behind a url: a web page's text, or an exportable Tako card's data rows. Batch up to 10 urls in one call — each one is billed and fails on its own. Fetch only cards that `tako_search` marked `exportable: true`. Rows bill per 1,000 delivered, so set `max_rows` when the recent rows are enough. If a page is long, such as a filing or an annual report, set `query` to get back only the passages that match. Best for: reading one source in full — a page you need to quote, or the rows behind a card you need to compute over.
  - after: Fetch the full content behind a url: a web page's text, or an exportable Tako card's data rows. Batch up to 10 urls in one call — each one is billed and fails on its own. Fetch only cards that `tako_search` marked `exportable: true`. Rows bill per 1,000 delivered, so set `max_rows` when the recent rows are enough. If a page is long, such as a filing or an annual report, set `query` to a question to get back only the passages that answer it. Best for: reading one source in full — a page you need to quote, or the rows behind a card you need to compute over.

Full history: https://api.protogrid.dev/v1/servers/io.github.TakoData%2Ftako-mcp/changes

## Remotes

- https://mcp.tako.com/mcp (streamable-http, auth none, reachable true, uptime30d 1)

## Packages

_none_

## Tools (4)

- `tako_available_data`: Find what data Tako holds on an entity or a metric, and the canonical name it holds it under.

Best for: coverage questions themselves, or a metric's canonical name before a priced search. Put a company, person, or place in `q` to list the metrics tracked on it; put a metric in `q` to list the entities it covers. Add `metric` when you know the measure — you get the resolved pair and a ready-to-run `next_call`.

Search on the canonical names it returns, not your own phrasing — `tako_search` matches the graph's names, not yours. A name here means the graph tracks it, not that a card exists. If the follow-up search comes back empty, say Tako has no card for it rather than rephrasing the query. Hand a node id to `tako_graph_related` to see what else connects to it.
- `tako_contents`: Fetch the full content behind a url: a web page's text, or an exportable Tako card's data rows. Batch up to 10 urls in one call — each one is billed and fails on its own.

Fetch only cards that `tako_search` marked `exportable: true`. Rows bill per 1,000 delivered, so set `max_rows` when the recent rows are enough. A website-traffic or SEO card also bills a per-unit data minimum that `max_rows` doesn't bound. If a page is long, such as a filing or an annual report, set `query` to a question to get back only the passages that answer it.

Best for: reading one source in full — a page you need to quote, or the rows behind a card you need to compute over.
- `tako_graph_related`: Explore what a graph node connects to — its metrics, the entities a metric covers, competitors, industry, index membership, and sources.

Two modes. Pass `node_id` alone for the map: every relation group with its key, label, total, and its first 3 names. Pass `relation` to page one group, where each item comes back with the id that explores it. Read a key off the map rather than guessing — an unknown key returns an empty group, not an error.

Best for: expanding a node you already resolved. Resolve a name to a node id with `tako_available_data` first. A metric listed here means the graph tracks it, not that a card exists — `tako_search` is the final check.
- `tako_search`: Search Tako's data graph and the live web in one call: many results at once, as structured cards plus web results, with the top card rendered inline as a chart.

It finds data; `tako_contents` fetches it. Each card carries a headline value, node ids, and a url — pass the url to `tako_contents` for rows (`exportable: true` cards) or a web result's full page text. When `exportable` is false the rows are locked — read the headline value from the card's `description`. Website-traffic and SEO cards bill a per-unit data minimum above the search price.

Best for: breadth — fan out several narrow queries in parallel. Each query resolves one metric — for one entity, or a comparison set ("Apple revenue", "Nvidia vs AMD gross margin"); several metrics or topics in one query retrieve poorly. To learn what Tako covers, or a metric's canonical name, run `tako_available_data` first, then search on the canonical name it returns.