Skip to contentATP Token
Developer gateway

Point base_url at ATP.
Change nothing else.

OpenAI, Anthropic and Gemini surfaces all run through one gateway. Request logs and credits land in one console.

Get started freeView docs
quickstart.py · OpenAI SDK
from openai import OpenAI

client = OpenAI(
  base_url="https://api.atptoken.ai/v1",
  api_key="atp-•••"
)

r = client.chat.completions.create(
  model="<model from /v1/models>",
  messages=[{"role":"user","content":"hi"}],
)
Agent skills

Let your agent wire itself up

Install the ATP skills once and your coding agent already knows the auth, the model ids, the payloads and the async media flow — no provider SDKs to explain.

Run it in your project

curl -fsSL https://atptoken.ai/skills/install.sh | sh -s -- claude

Skills are Markdown only — they never carry your key.

Or hand this to the agent

Install the ATP Token agent skills by running:
curl -fsSL https://atptoken.ai/skills/install.sh | sh -s -- claude

Then read the skills and call the ATP gateway at https://api.atptoken.ai/v1 using the atp- key in my environment.

Paste it into Claude Code, Codex or any agent that reads SKILL.md.

Skills reference
/v1/models
gpt-5.4-proclaude-fable-5gpt-5.6-solgpt-5.5claude-opus-4-8claude-opus-4-7claude-sonnet-5gpt-5.6-terraclaude-sonnet-4-6gpt-5.4gemini-3-1-pro-previewgemini-3-5-flashqwen-3-6-maxqwen-3-8-maxqwen-3-7-maxqwen-3-7-plusqwen-3-7-flashqwen-3-6-plusqwen-3-6-flashqwen-plus-charactergpt-5.6-lunaclaude-haiku-4-5glm-5-1kimi-k2.6glm-5-2kimi-k2.7-codekimi-k3deepseek-v4-proglm-5gemini-3-flash-previewdeepseek-v4-flashdeepseek-v3-2gpt-image-2seedance-2-0seedance-2-0-fastfish-s2-protext-embedding-3-smallkling-v3-standardkling-v3-standard-i2vkling-v3-prokling-v3-pro-i2vkling-o3-standardkling-o3-standard-i2vkling-o3-standard-referencekling-o3-standard-reference-7kling-o3-prokling-o3-pro-i2vkling-o3-pro-referencekling-o3-pro-reference-7kling-o3-standard-v2vkling-o3-standard-video-editkling-o3-pro-v2vkling-o3-pro-video-editwan-2-7-t2vwan-2-7-i2vhappyhorse-1.1-t2vhappyhorse-1.1-i2vhappyhorse-1.1-r2vhappyhorse-1.0-video-editnano-banananano-banana-editnano-banana-pronano-banana-pro-editnano-banana-2qwen-image-3.0qwen-image-3.0-pronano-banana-2-editgemini-3-5-flash-litegemini-3-6-flashgpt-5.4-proclaude-fable-5gpt-5.6-solgpt-5.5claude-opus-4-8claude-opus-4-7claude-sonnet-5gpt-5.6-terraclaude-sonnet-4-6gpt-5.4gemini-3-1-pro-previewgemini-3-5-flashqwen-3-6-maxqwen-3-8-maxqwen-3-7-maxqwen-3-7-plusqwen-3-7-flashqwen-3-6-plusqwen-3-6-flashqwen-plus-charactergpt-5.6-lunaclaude-haiku-4-5glm-5-1kimi-k2.6glm-5-2kimi-k2.7-codekimi-k3deepseek-v4-proglm-5gemini-3-flash-previewdeepseek-v4-flashdeepseek-v3-2gpt-image-2seedance-2-0seedance-2-0-fastfish-s2-protext-embedding-3-smallkling-v3-standardkling-v3-standard-i2vkling-v3-prokling-v3-pro-i2vkling-o3-standardkling-o3-standard-i2vkling-o3-standard-referencekling-o3-standard-reference-7kling-o3-prokling-o3-pro-i2vkling-o3-pro-referencekling-o3-pro-reference-7kling-o3-standard-v2vkling-o3-standard-video-editkling-o3-pro-v2vkling-o3-pro-video-editwan-2-7-t2vwan-2-7-i2vhappyhorse-1.1-t2vhappyhorse-1.1-i2vhappyhorse-1.1-r2vhappyhorse-1.0-video-editnano-banananano-banana-editnano-banana-pronano-banana-pro-editnano-banana-2qwen-image-3.0qwen-image-3.0-pronano-banana-2-editgemini-3-5-flash-litegemini-3-6-flash
69models behind one key
5modalities — text to video
3API surfaces, one wallet
$5min top-up, never expires
Models

Three API surfaces, one gateway, every model.

Call the surface your SDK already speaks. The gateway routes it to the provider pool for that model, keeps the key scoped to its project, and meters every request.

Three surfaces, one key

  • OpenAI SDK
    /v1/chat/completionsclient.chat.completions.create(model, messages)
  • Anthropic Messages
    /v1/messagesclient.messages.create(model, max_tokens, messages)
  • Gemini-style
    /v1beta/models/…:generateContentgenerateContent(model, contents)

OpenAI-compatible

Keep the SDK shape you already use; only the base URL changes.

Project-scoped

Each key can only reach the models enabled on its project.

Fallback-ready

Requests route through the provider pool configured for that model.

Discoverable

A key can ask the gateway which models it may call.

How it works

From zero to first request in three steps.

Observability

Governance and billing, built into the request path.

  • Per-request logs — status, model, tokens, credits
  • Keys scoped to one project, models allow-listed per key
  • One wallet across providers — spend is visible before the invoice
Open your console
requests / min · simulated
120
req/min
p50 412ms30s windowlive lerp 8%
Get started

One key. Every model.

Change one base URL today. Your first request costs less than a cent.

Create accountRead the docs