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

# Threads data for AI agents

> Threads posts and replies, user profiles, posts, replies and reposts, and search for top or recent content and profiles through AgentKey — no Meta API needed.

Threads is Meta's text-first network and the closest thing to a second X for
public conversation, yet its official API only covers accounts you own and the
web app shows little without a login. Through AgentKey your agent reads public
Threads data the same way it reads everything else — `find_tools`,
`describe_tool`, `execute_tool` — and the same prompts that work for X work
here.

## What you can ask

* *"What has @zuck posted on Threads lately?"*
* *"Pull the replies on this Threads post and summarize the reaction."*
* *"Search Threads for recent posts about the new iPhone."*
* *"Find Threads profiles related to climate journalism."*
* *"Show me this user's reposts and replies, not just original posts."*
* *"What are the top posts about 'AgentKey' on Threads?"*
* *"Give me the text and media of this Threads post from its URL."*
* *"Compare how these two accounts talk about the same product launch."*

## What's covered

**Posts**

* Post detail by ID — text, media and author
* Post detail straight from a copied URL, so a shared link needs no parsing
* The comments (replies) under a post

**Users**

* Profile by username, which also yields the numeric user ID
* Profile by numeric user ID, for accounts you already resolved
* A user's own posts
* Their replies to other people, as a separate stream
* Their reposts

**Search**

* Top content for a keyword — what Threads ranks highest
* Recent content for a keyword — newest first
* Profile search by name or keyword
* Results carry the post and user identifiers the detail tools above take, so
  search → detail → replies needs no manual ID extraction

A typical task starts with search (content or profiles), resolves the account,
then pulls its posts, replies and reposts, and finally reads the replies under
the posts that matter. Each step is a separate call, so budget for the number
of accounts and posts you plan to walk.

## How to use it

Pass the request in plain language and let the router pick the tool:

```text theme={null}
find_tools(q="recent Threads posts about the new iPhone")
find_tools(prefix="social/threads")        # browse everything for Threads
describe_tool("<returned Provider/Operation>")
execute_tool("<Provider/Operation>", { ...params })
```

Both English and Chinese phrasing work, and the platform's usual English and Chinese names all land here. Always `describe_tool` before executing —
parameters and per-call costs are listed there, and the user tools differ in
whether they take a username or a numeric ID.

> [!NOTE]
>
> **Building with code?** Threads is available through the catalog tools above
> only; there are no fixed-contract Scenario Methods for it yet. See the
> [Scenario Methods overview](/api-reference/scenario/overview) for the
> platforms that have them.

## FAQ

> **Do I need a Threads account or a Meta developer app?**
>
> No. AgentKey holds the upstream access; your agent only needs your AgentKey
> subscription.

> **Can it read private profiles?**
>
> No. Only public posts, replies, reposts and profiles are available, and
> every tool is read-only — nothing is posted or followed on your behalf.

> **How fresh is the data?**
>
> Every call goes to the platform at request time; nothing is served from a
> stale archive. "Recent" and "top" search reflect what Threads returns at
> that moment.


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