API & AI agents: drive Pamba programmatically

Everything the Pamba UI does, an agent can do through the API. The full public API reference lives at pamba.app/api: endpoint auth, request and response structures, credit usage, and examples.

Authentication

• •Base URL: https://api.pamba.app
• •Agent auth: X-API-Key header. Create keys in workspace settings (open the workspace switcher, then Settings).
• •JWT callers must include X-Workspace-Id on workspace-scoped routes; API-key callers don't, since the key already carries the workspace.

Give your agent the guide

Download the guide matching your agent and drop it in its context:

• •CLAUDE.md guide for Claude agents
• •Codex guide
• •Generic agent guide

Once your agent has it, a prompt like "What can you do with Pamba?" should produce a concrete menu of workflows, not stale endpoint guesses.

What your agent can do

Track TikTok accounts and mine analyzed videos; add ideas in bulk and remix scripts; create fully automated project batches; generate clips and select takes; schedule, reschedule, cancel, and retry posts; configure warming; create avatars from reference images; manage starting-frame libraries; and pull analytics dashboards and account metrics.

Credits

Generation spends credits (clip generation per model-second, avatar photos and scenes at 15 credits per image, voices at 100, premium voice clones at 1500, hook purchases at 200). Failed operations are never charged, and insufficient-credit requests return 402 Payment Required with remaining_credits and required_credits. The live price list is public at GET /usage/pricing; use GET /user for subscription state and credit totals.

This help center is agent-readable too

Every article here is available as raw markdown: see /llms.txt for the index and /llms-full.txt for the whole corpus in one document.