Use Pika
from your agent

Give your agent a creative engine: generate video, image, audio, and speech through a single Pika API.

Send this to ChatGPT
You can call the Pika API to generate video, image, audio, music, and speech. Base URL: https://089e99349ace.pikalabs.app Auth header: X-API-Key: pk_live_YOUR_KEY Recommended models: - video.text-to-video pika/pika-2.5/text-to-video - video.image-to-video pika/pika-2.5/image-to-video - video.cinematic bytedance/seedance-2.0/text-to-video - image.generate openai/gpt-image-2/text-to-image - image.edit google/gemini-3.1-flash-image/image-to-image - video.animate kling/kling-v3/standard/image-to-video Full model list + every endpoint's request/response spec: fetch https://dev.pikalabs.app/llms.txt (each model there links to /llms/{model} for its complete spec). Calling convention: 1. POST /v1/media/{api_id} with that model's JSON body 2. Poll GET /v1/media/jobs/{request_id} until status == "completed" 3. Fetch GET /v1/media/jobs/{request_id}/content for the result url When the user asks for a video, default to seedance-2.0 at 720p, 5s. Always confirm the model and resolution before spending more than 100 credits.
OR

Test the API yourself with the quick-start guide

Don’t have an account? Sign up — it’s free

All about the Pika API

Everything teams ask before integrating.

The API is self-serve — sign in, create an API key, and you get access to every Pika model listed above along with a quickstart guide. Most teams are running their first generation within the hour. Questions about volume or custom needs? Reach out via the contact form below.

Every API customer gets a shared Slack channel with our developer-relations team. For non-urgent questions, email api@pika.art and we typically respond within one business day.

Pricing is per-second of generated video and varies by model and resolution. Most models are $0.04/sec at 720p and $0.06/sec at 1080p. Volume discounts kick in automatically above 10,000 seconds per month — contact us for an enterprise quote.

Default accounts get 60 requests per minute and 10 concurrent jobs. Limits scale automatically with usage; enterprise plans get dedicated capacity.

Yes. All API outputs are licensed for commercial use under your Pika API agreement. You retain ownership of your prompts and inputs.

Get started with the Pika API

The Pika API is self-serve — sign in to create an API key and start building in minutes.

Start building

Questions about models, pricing, or volume? Tell us about your product and we'll get back to you within one business day.