API documentation

API Many exposes every model through one base URL and one key. Text endpoints follow the OpenAI and Anthropic formats, so official SDKs work unchanged; image, video and music run as asynchronous tasks.

Base URL https://apimany.com/v1

Quickstart

Sign in and create a key under API keys.
Add credit from $10 under Credits.
Send your first request:

curl https://apimany.com/v1/chat/completions \
  -H "Authorization: Bearer $APIMANY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "claude-sonnet-5-5", "messages": [{"role": "user", "content": "Hello"}]}'

Authentication

Send your key in the Authorization header as a Bearer token. Keys are secret: call the API from your server, never from a browser or mobile app.

Authorization: Bearer sk_apimany_xxxxxxxxxxxxxxxx

# Anthropic-style clients may send the key as:
x-api-key: sk_apimany_xxxxxxxxxxxxxxxx

SDKs

Use the official OpenAI SDK (Python, Node.js, Go, Java, .NET) with base_url https://apimany.com/v1, or the official Anthropic SDK with base_url https://apimany.com. Any HTTP client works for media endpoints.

Billing

Every request reserves its maximum cost from your balance when it starts and is settled when it finishes:

Text: billed by the input, cached and output tokens the model reports, at the per-million prices on its page.
Images: billed per image at the chosen resolution or quality; token-priced image models settle on reported usage.
Video: billed per second at the chosen resolution (or per clip for fixed-length models).
Failed, rejected and timed-out requests are not charged: the reserve is returned in full.

Send an Idempotency-Key header (1–200 characters) to make retries safe: a repeated key returns the original record instead of running again.

Chat completions POST/v1/chat/completions

OpenAI-compatible chat for every language model, including Claude, Gemini, DeepSeek, Grok, Qwen and Kimi. Streaming, tools, JSON mode and image inputs follow the OpenAI format. Streamed responses include usage in the final chunk.

curl https://apimany.com/v1/chat/completions \
  -H "Authorization: Bearer $APIMANY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "deepseek-v4.1-flash",
    "stream": true,
    "messages": [
      {"role": "system", "content": "You are concise."},
      {"role": "user", "content": "Summarize HTTP/3 in two sentences."}
    ]
  }'

Anthropic Messages POST/v1/messages

Claude models also accept the native Anthropic Messages API, including streaming, tool use, extended thinking and prompt caching. The anthropic-version and anthropic-beta headers are forwarded.

import anthropic

client = anthropic.Anthropic(base_url="https://apimany.com", api_key="YOUR_API_KEY")
message = client.messages.create(
    model="claude-opus-5-5",
    max_tokens=1024,
    messages=[{"role": "user", "content": "Write a haiku about APIs."}],
)
print(message.content[0].text)

Responses POST/v1/responses

The OpenAI Responses API for models that support it.

curl https://apimany.com/v1/responses \
  -H "Authorization: Bearer $APIMANY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "gpt-6.1-sol", "input": "Explain recursion to a child."}'

Embeddings POST/v1/embeddings

curl https://apimany.com/v1/embeddings \
  -H "Authorization: Bearer $APIMANY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "text-embedding-3-small", "input": ["first text", "second text"]}'

Images POST/v1/images/generations

Text-to-image and image editing. Each model documents its own fields (size, resolution, quality, reference images) on its model page; send them as listed there. The call returns a task immediately:

curl https://apimany.com/v1/images/generations \
  -H "Authorization: Bearer $APIMANY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-image-2",
    "prompt": "A watercolor fox in a snowy forest",
    "size": "16:9",
    "resolution": "2k",
    "image_urls": ["https://example.com/style-reference.png"]
  }'

{
  "id": "2c7f9a8e-0d1b-4b5e-9a5d-6f0c1e2d3a4b",
  "object": "generation.task",
  "model": "seedance-2.5",
  "status": "pending",
  "reserved_usd": 2.16,
  "task_id": "2c7f9a8e-0d1b-4b5e-9a5d-6f0c1e2d3a4b"
}

Videos POST/v1/videos/generations

Text-to-video, image-to-video, references, edits and extensions — depending on the model. Per-second models reserve the requested duration and settle on the delivered clip.

curl https://apimany.com/v1/videos/generations \
  -H "Authorization: Bearer $APIMANY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "seedance-2.5",
    "prompt": "Slow dolly-in on a neon alley after rain",
    "resolution": "720p",
    "duration": 8,
    "size": "16:9",
    "callback_url": "https://your-app.com/webhooks/apimany"
  }'

Music POST/v1/music/generations

curl https://apimany.com/v1/music/generations \
  -H "Authorization: Bearer $APIMANY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model": "suno", "custom": false, "version": "v6", "prompt": "Late-night lo-fi piano with soft rain"}'

Tasks & polling GET/v1/tasks/{id}

Poll every 3–10 seconds until status is completed or failed. Statuses: pending, processing, completed, failed. Completed tasks list signed file URLs that stay valid for 30 days; download anything you need to keep.

curl https://apimany.com/v1/tasks/TASK_ID -H "Authorization: Bearer $APIMANY_API_KEY"

Webhooks

Add callback_url (public HTTPS) to any media request and we POST the finished task to it, retrying with backoff for up to six attempts. Verify X-Apimany-Signature with the webhook secret from your API keys page.

import crypto from 'node:crypto'

// Express-style handler; keep the raw body for the signature check.
export function verify(rawBody, headers, secret) {
  const timestamp = headers['x-apimany-timestamp']
  const expected = crypto.createHmac('sha256', secret.replace(/^whsec_/, ''))
    .update(`${timestamp}.${rawBody}`).digest('hex')
  return crypto.timingSafeEqual(Buffer.from(expected), Buffer.from(headers['x-apimany-signature']))
}

// Payload: { "event": "task.completed" | "task.failed", "data": { ...task, "record": { ... } } }

Transcription POST/v1/audio/transcriptions

OpenAI-compatible speech-to-text with multipart upload (up to 25 MB).

curl https://apimany.com/v1/audio/transcriptions \
  -H "Authorization: Bearer $APIMANY_API_KEY" \
  -F model=gpt-4o-transcribe \
  -F [email protected] \
  -F response_format=json

Models GET/v1/models

Public list of every model with type, endpoint and pricing; filter with ?kind=chat|image|video|audio|embedding.

# Every model with its pricing (no key needed)
curl https://apimany.com/v1/models

# Only video models
curl "https://apimany.com/v1/models?kind=video"

# One model with its parameters
curl https://apimany.com/v1/models/seedance-2.5

Records GET/v1/records

Every call you make — text, media and audio — has a record with status, usage, cost and outputs. Filter by model, kind, status, api_key_id or since.

curl "https://apimany.com/v1/records?limit=20&kind=video&status=completed" \
  -H "Authorization: Bearer $APIMANY_API_KEY"

curl https://apimany.com/v1/records/RECORD_ID -H "Authorization: Bearer $APIMANY_API_KEY"

Balance GET/v1/account/balance

curl https://apimany.com/v1/account/balance -H "Authorization: Bearer $APIMANY_API_KEY"
# { "object": "balance", "currency": "usd", "balance_usd": 42.17, "debt_usd": 0 }

Errors

Errors use the OpenAI error shape, so SDK exceptions work as usual.

{
  "error": {
    "message": "Your balance is too low for this request. Top up and try again.",
    "type": "insufficient_quota",
    "param": null,
    "code": "insufficient_balance"
  }
}

HTTP	code	Meaning
400	invalid_request	A parameter is missing or not allowed for this model. Nothing was charged.
401	invalid_api_key	The key is missing, malformed or revoked.
402	insufficient_balance	Your balance cannot cover the reserve for this request.
404	model_not_found	The model id does not exist. List models with GET /v1/models.
409	idempotency_replay	This Idempotency-Key was already used; fetch the earlier record instead.
413	request_too_large	The request body is over the size limit. Send media as URLs.
429	rate_limited	Too many requests per minute for this key. Retry with backoff.
502	generation_failed	Every channel for this model failed. Nothing was charged; retry shortly.
503	model_unavailable	The model is temporarily unavailable.

Limits

Request body: up to 25 MB for text and 40 MB for media requests. Pass large files as HTTPS URLs.
Playground uploads: images up to 10 MB, video up to 100 MB, audio up to 50 MB.
Rate limit: 120 requests per minute per account by default. Contact us for higher limits.
Generated files are kept for 30 days. Tasks that never finish are reviewed and refunded.
Model capabilities and per-model limits follow the model maker.