Meta Threads & Truth Social Real-Time MCP Server for AI Agents
Connect Claude Code, Cursor, and real-time news or trading agents to Meta Threads profiles, discussion feeds, and keyword search alongside low-latency Truth Social executive and political posts.
claude mcp add --transport http agentapigate-threads "https://mcp.agentapigate.ai/mcp/custom?tools=threads.user.profile,threads.user.posts,threads.posts.search,truthsocial.user.profile,truthsocial.user.posts" \ --header "Authorization: Bearer aag_live_YOUR_KEY"
# 1. Inspect deterministic 402 quote (Base USDC + Solana USDC + Tempo MPP): curl -i "https://api.agentapigate.ai/v1/threads/posts/search?query=openai&limit=12&format=compact" # 2. Or execute immediately with an aag_live_* / scoped aag_agent_* key: curl -s "https://api.agentapigate.ai/v1/threads/posts/search?query=openai&limit=12&format=compact" \ -H "Authorization: Bearer $AAG_API_KEY" \ -H "X-AAG-Max-Cost: 0.0100"
{
"ok": true,
"meta": {
"endpoint": "threads.posts.search",
"format": "compact",
"cost_usd": 0.003,
"tokens_est": 1870,
"raw_tokens_est": 86535,
"tokens_saved_pct": 98,
"latency_ms": 4692,
"content_trust": "external_untrusted"
},
"data": {
"query": "openai",
"count": 12,
"items": [
{
"id": "4034063519339695094",
"author": "financialtimes",
"text": "Anthropic's IPO has transformed the personal fortunes of its seven co-founders, who left OpenAI in 2021 after becoming frustrated with Sam Altman's leadership. https://ft.trib.al/2cBStuF",
"likes_count": 443,
"replies_count": 18,
"published_at": "2026-11-19T19:51:17.000Z",
"url": "https://www.threads.net/@financialtimes/post/Df732W_msv2"
},
{
"id": "4034323169053444273",
"author": "andikaluis",
"text": "HE DOESN'T TRUST ONE AI WITH HIS MONEY. HE MAKES 3 AI AGENTS VOTE. 🤯\n\nReported profit today: +$4,631.\n\nClaude Opus, Jev, and OpenAI Dots each analyze the same trading opportunity.\n\nNo trade gets approved until all three agree.\n\nMeanwhile, another agent monitors whale activity across major exchanges, looking for unusual movements and potential opportunities.\n\nOne AI finds the setup. Three agents challenge the decision. Risk controls stand between the signal and execution.",
"likes_count": 4,
"replies_count": 1,
"published_at": "2026-11-20T04:27:25.000Z",
"url": "https://www.threads.net/@andikaluis/post/Df8y4wojwix"
},
{
"id": "4034239155374264614",
"author": "nytimes",
"text": "The seven people who left OpenAI to create Anthropic in 2021 now each hold stakes worth tens of billions of dollars after the company made its stock market debut. https://trib.al/VsW9YKg",
"likes_count": 101,
"replies_count": 125,
"published_at": "2026-11-20T01:40:16.000Z",
"url": "https://www.threads.net/@nytimes/post/Df8fyMzDhEm"
},
{
"id": "4034606759066862918",
"author": "cmdenter",
"text": "OpenRouter: 20+ free models through a single API key — no credit card. 20 RPM, 50/day (1,000/day with $10 top-up). Up to 1M-token context on some models.\n\nThe real value is failover: route across 70+ providers through one OpenAI-compatible endpoint. If Groq throttles you, the request falls back to Cerebras.\n\nhttps://openrouter.ai\n\nCmdEnter AI · AI moves fast. Stay ahead. 🤖",
"likes_count": 0,
"replies_count": 0,
"published_at": "2026-11-20T13:50:27.000Z",
"url": "https://www.threads.net/@cmdenter/post/Df9zXiakgVG"
},
{
"id": "4034735884129053335",
"author": "leoisadev",
"text": "OpenAI needs another code red\n\nThey're falling behind with:\n- products (codex app is super laggy)\n- models aren't good\n- subscription doesn't give much value\n\nI could be wrong and they could launch something insanely big but I have a feeling until next year we won't get any new good models from them.",
"likes_count": 1,
"replies_count": 0,
"published_at": "2026-11-20T18:07:07.000Z",
"url": "https://www.threads.net/@leoisadev/post/Df-QujgCKqX"
},
{
"id": "4034003889766564892",
"author": "tovelie.tech",
"text": "At OpenAI, performance alerts now trigger Codex to investigate regressions and propose fixes for engineer review. After deployment, another agent checks whether performance returned to baseline. OpenAI says tWhy Engineers Mount This Connector via AgentAPIGate
- 01 · GUARANTEE
Full Text-Social Coverage Alongside X & Bluesky: cross-verify breaking tech launches, policy statements, and brand sentiment across Threads, Truth Social, X, and Bluesky in one schema.
- 02 · GUARANTEE
Zero WAF / Challenge Blocks: Truth Social and Threads block raw datacenter fetches; AgentAPIGate delivers clean normalized JSON and Markdown in ~700ms p50.
- 03 · GUARANTEE
Deterministic Pay-Per-Call Economics: $0.002–$0.004/call with zero monthly seat minimums, X-AAG-Cost-USD receipt headers, and Base/Solana USDC x402 + Tempo MPP support.
Included MCP Tools & Deterministic Per-Call Pricing
Browse all 136 endpoints →| MCP Tool / Endpoint ID | Summary | API Key Price | x402 / MPP | p50 Latency | Bonus Credit |
|---|---|---|---|---|---|
| threads_user_profile threads.user.profile | Fetch public Meta Threads profile details including username, full name, biography, follower count, verification flag, and bio links. | $0.0020 | $0.0020 | 1561 ms | Eligible |
| threads_user_posts threads.user.posts | Retrieve recent public Threads posts by username with full text, like/reply/repost counts, and pagination cursor. | $0.0020 | $0.0020 | 5569 ms | Eligible |
| threads_posts_search threads.posts.search | Search public Meta Threads discussions by keyword or topic with author handles, engagement counts, and post URLs. | $0.0030 | $0.0030 | 4692 ms | Eligible |
| truthsocial_user_profile truthsocial.user.profile | Retrieve public Truth Social account metadata, follower/following counts, Truths count, bio, and join timestamp. | $0.0040 | $0.0040 | 336 ms | paid_usd / x402 |
| truthsocial_user_posts truthsocial.user.posts | Fetch latest public Truth Social posts ("Truths"), reblogs, media attachments, and engagement metrics for real-time news and macro trading agents. | $0.0040 | $0.0040 | 2408 ms | paid_usd / x402 |
Frequently Asked Questions
How fast are Truth Social posts refreshed for event-driven trading and news agents?
truthsocial.user.posts uses a short 120-second edge cache TTL (and supports standard cache headers) so macro sentiment and prediction-market agents capture new posts with sub-second gateway overhead.
Does Threads search support pagination across multi-page topic threads?
Yes. Both threads.user.posts and threads.posts.search return next_cursor so agents can page deeper only when initial results warrant further inspection.