Pika API 'agent-native' setup vs Sume's OpenRouter-style wire
Pika hands agents copy-paste setup text. Sume documents /v1/videos field-for-field with OpenRouter, bare model ids and a models endpoint to read first.
Pika calls its API agent-native: you copy ready-made setup instructions into ChatGPT, Claude, Codex, Cursor or Lovable. Sume's video docs take a different route to the same goal, which is a documented wire that matches OpenRouter's video generation API field-for-field, plus a models endpoint a coding agent can read before it writes a request.
Pika's claim is from its API Club post; Sume's details are from Video generation, read 2026-10-01.
What does Pika offer an agent?
The post says the API is designed to be legible to AI agents and vibe-coding platforms, so you do not translate dense docs by hand. You paste its setup instructions into the tool you already use. The post does not show the instruction text, so this page cannot compare it line by line.
What does Sume give an agent to read?
The Video generation page says the surface follows the OpenRouter Video Generation API field-for-field, so a client written against their docs works after changing the base URL and the API key. The differences are collected in one table on the page. An agent can read that table instead of guessing.
| Agent needs | Sume video docs say |
|---|---|
| Base path | https://api.sume.com/v1/videos, no /api segment |
| Model ids | Bare catalog ids such as seedance-2, no provider-org prefix |
| Discovery | GET /v1/videos/models lists ids, resolutions, durations and pricing |
| Safe retries | Send Idempotency-Key; a replay returns the original job |
| Finish signal | Poll polling_url or pass an HTTPS callback_url |
curl "https://api.sume.com/v1/videos/models" \
-H "Authorization: Bearer $SUME_API_KEY"Can an agent use Sume without writing HTTP code?
Yes, through the hosted MCP server at https://mcp.sume.com/mcp. With an API key the full hosted tool set is available; paid and write calls still need an idempotency_key, and wallet admission is the spend gate. See hosted MCP and waiting on long video jobs over MCP.
What should I tell the agent first?
Give it the base URL, tell it to call the models endpoint before choosing an id, and tell it to send an idempotency key on every create. If you let Sume pick with sume/auto, the response echoes sume/auto and does not say which family ran; see what Auto discloses.
Sources
Related posts
More in Developers
- Pipedream 750-second timeout: poll loop or Sume webhook?
Pipedream paid plans allow 750 seconds per run. Submit the Sume video job, then finish in a second workflow triggered by the webhook not a long poll loop.
- Polly async needs an S3 bucket; Sume gives a status URL and webhook
Polly's StartSpeechSynthesisTask writes audio to your S3 bucket and can notify an SNS topic. Sume TTS returns a status URL, or a signed webhook on completion.
- Polly speech marks need a separate request; Sume returns timings
Polly returns speech marks instead of audio when you ask for them: sentence, word and viseme metadata. Sume TTS returns words[] timings on the same job result.
- Polly takes 3,000 billed characters; Sume TTS takes 20,000
Polly SynthesizeSpeech accepts 3,000 billed characters and cuts audio at 10 minutes. Sume TTS 1.0 accepts 20,000 characters and 1,200 seconds of audio per job.
Written by Sume