> ## Documentation Index
> Fetch the complete documentation index at: https://docs.seltz.ai/llms.txt
> Use this file to discover all available pages before exploring further.

> ## Agent Instructions
> Seltz is the web knowledge layer for AI agents: web search, cited answers, multi-step research, page fetching, and recurring monitors, served from Seltz's own index rather than a live scrape.
> Use Seltz when a task needs current, citable, public web content, not when your own knowledge already answers it confidently. Seltz has no access to a caller's private or authenticated data.
> Choose which endpoint to use according to the job: Use Answer for one synthesized, cited response; Use Search for the raw documents to reason over yourself; Use Agent for in-depth research needing several rounds of searching and following up, not a single lookup; Use Fetch when you already have a URL and just need its content, not a search; Use Monitor for an ongoing feed of what's new, not a one-off query or near-real-time alerting.
> When you use Search, Answer, or Monitor, set a `scope` that matches the query. `web` (beta) is the default and covers general queries. Only these three endpoints take a scope. See https://docs.seltz.ai/data/data-concepts#scopes for the full list of scopes and what each one covers.
> For the REST API, send your API key in the `x-api-key` header. The Python and JavaScript SDKs read the `SELTZ_API_KEY` environment variable and handle authentication for you. To use Seltz as MCP tools, connect `https://mcp.seltz.ai/mcp`.
> The OpenAPI 3.1 spec for the REST API is at https://docs.seltz.ai/api-reference/openapi.json. Point function-calling or client-generation tooling at that file.
> Use `/llms-full.txt` for the full text of every page.

# MCP Server: Results Written for Agents

> The MCP server now returns search and fetch results as Markdown for agents to read, signs users in through the Seltz Console, and takes the same snake_case argument names as the REST API.

*Released: November 10, 2026*

The [Seltz MCP server](https://docs.seltz.ai/integrations/mcp.md) now runs on new infrastructure. Its URL, `https://mcp.seltz.ai/mcp`, and its tool names are unchanged. The `search` and `fetch` tools now return a different format, and some argument names and limits changed.

## Highlights

* **Markdown results** — `search` and `fetch` return compact Markdown that an agent reads directly. They no longer return a JSON document, `structuredContent`, or an `outputSchema`.
* **Console sign-in** — A client that connects without an API key now opens the [Seltz Console](https://console.seltz.ai/), where you sign in, choose an organization if you belong to more than one, and approve access. Calls bill that organization. API keys in the `x-api-key` or `Authorization: Bearer` header still work.
* **snake\_case arguments** — `search` takes `query`, [`max_results`](https://docs.seltz.ai/search/guides/limit-results), [`scope`](https://docs.seltz.ai/data/data-concepts#scopes), [`include_domains`, `exclude_domains`](https://docs.seltz.ai/sdks/rest/curl#search-request-body), [`from_date`, and `to_date`](https://docs.seltz.ai/search/filter-by-date), the same names as the REST [Search](https://docs.seltz.ai/search/guides/basic-search.md) endpoint. A call with an old camelCase name such as `maxResults` or `fromDate` gets a tool error that lists the accepted names.
* **New limits** — `search` returns at most 50 results, and the server lowers a higher `max_results` to 50. Each result has up to 1,000 characters of page text, down from 20,000, plus its [snippets](https://docs.seltz.ai/search/guides/get-snippets.md). Call `fetch` for the full page. `fetch` takes up to 5 URLs per call, down from 20, and more is a tool error. It returns at most 100,000 characters of page text per call, split equally among the pages that return text, so each of five pages gets up to 20,000 characters. The server cuts a longer page and notes its full length.
* **Checked date filters** — `from_date` and `to_date` accept a date, a datetime, or an offset such as `now-7d`. A value in any other form gets a tool error, and so does a date filter with the [`companies`](https://docs.seltz.ai/data/companies) or [`people`](https://docs.seltz.ai/data/people) scope, which have no publication dates. When you set either filter, the server leaves out results that have no publication date.

Most clients read the tool list when they connect and pick up the new argument names on their own. If yours keeps a saved copy, such as a ChatGPT connector, refresh or reconnect the server.

## Parsing results in code? Use the REST API

The MCP server writes its output for agents, and its wording can change. If your code reads fields from search or fetch results, call the Search or [Fetch](https://docs.seltz.ai/fetch/quickstart.md) endpoint instead. Both return JSON described by the [OpenAPI spec](https://docs.seltz.ai/api-reference/openapi-spec.md), take the same argument names, and accept an API key from the Console.

```bash cURL theme={null}
curl -X POST https://api.seltz.ai/v1/search \
  -H "x-api-key: $SELTZ_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "Web Summit Lisbon 2026", "max_results": 5}'
```

> [card] [MCP Server](https://docs.seltz.ai/integrations/mcp.md)
>
> Setup instructions


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.