captapi

Official typed Python SDK for the Captapi social media API — 178 endpoints across 32 platforms (YouTube, TikTok, Instagram, Facebook, X/Twitter, Reddit, Threads, Bluesky, Pinterest, LinkedIn, Twitch, Spotify, SoundCloud, GitHub, and more).

• Sync and async clients (Captapi / AsyncCaptapi), built on httpx
• Typed, documented method for every endpoint — great autocomplete for humans and AI coding agents
• Errors always raise CaptapiError with status code and detail — no silent failures
• One bounded retry on 502/503 (skipped when the API says the failure is persistent or already cached)

Install

pip install captapi

Quick start

from captapi import Captapi

client = Captapi(api_key="YOUR_KEY")  # or set CAPTAPI_API_KEY

transcript = client.youtube.transcript(url="https://youtube.com/watch?v=dQw4w9WgXcQ")
print(transcript["data"])

profile = client.tiktok.channel_details(url="https://tiktok.com/@nba")
posts = client.reddit.subreddit_posts(url="https://reddit.com/r/programming", limit=25)
ads = client.google_ad_library.company_ads(advertiser="nike.com")

Async

from captapi import AsyncCaptapi

async with AsyncCaptapi() as client:
    result = await client.instagram.channel_details(url="https://instagram.com/nba")

Error handling

Every failed request raises — nothing fails silently:

from captapi import Captapi, CaptapiError

try:
    client.instagram.channel_details(url="https://instagram.com/whoever")
except CaptapiError as err:
    print(err.status_code, err.code, err.detail)

Responses

All methods return the standard Captapi envelope as a dict:

{ "success": true, "data": { ... } }

Links

• Documentation
• Get an API key (100 free credits)
• TypeScript SDK: npm install @captapi/sdk
• MCP server for AI agents: @captapi/mcp
• CLI: @captapi/cli