Pricing
Pay-as-you-go, no plans or minimums. One flat rate for everything — conversation, inline edits, and native tool-use for IDE agents — or license private deployment.
Pay-as-you-go
25% off — pay as you go
One key, one rate for everything — conversation and inline edits, plus native tool-calling for autonomous IDE agents like Cursor Agent, Cline, Codex, and Claude Code.
- Claude Opus & Sonnet via one endpoint
- OpenAI Chat Completions + Responses
- Anthropic Messages API (Claude Code)
- Native tool-use / function calling included
- Streaming (SSE) responses
- Flat 25% off list, both input & output
- No tiers, seats, or minimums
Private Deployment
Licensed software
Licensed software you run in your own environment.
- Same universal multi-format API
- Runs inside your infrastructure
- Connect your own MCP tool servers
- Commercial license & support
- Onboarding & architecture review
- Volume pricing & SLA available
Per-token rates
One simple price per model — a flat 0.75× our list rate (25% off) on both input and output, with native tool-use included. Same rate for chat and IDE agents. The rate shows on each key in the console. No tiers, pay only for what you use.
| Model | List / 1M (in / out) | Your rate 0.75× (in / out) |
|---|---|---|
| Claude Opus | $5.00 / $25.00 | $3.75 / $18.75 |
| Claude Sonnet | $3.00 / $15.00 | $2.25 / $11.25 |
Prices per 1M tokens — list rate struck through, then the flat 0.75× rate you pay. The rate appears on each key in the console. The smart model is billed at the rate of the model it selects. Deeper multipliers available at volume — see all models.
Frequently Asked Questions
How do I get an API key?
Sign up in the customer console and create a key. Point any supported SDK at our base URL and you're live. One key covers everything — chat, inline edits, and native tool-use. Pay-as-you-go, no tiers or minimums.
What's included?
Everything, on one key. The same Claude Opus and Sonnet models over the same universal API — conversational chat and inline edits, plus native tool-calling so autonomous IDE agents (Cursor Agent, Cline, Codex, Claude Code) read files, write files, and run commands end-to-end. When you send tool definitions, you get native tool-use automatically — there's nothing to switch on.
Is this a subscription or plan?
Neither — it's pure pay-as-you-go. There are no monthly plans, seats, or minimums. You top up a balance and draw it down per token at a flat 0.75× of our list rate.
Do I need a different key for IDE agents?
No. One key does everything. Native tool-use is included on every account — send your tool definitions and Drawbridge returns native tool-calls, so IDE agents like Cursor Agent, Cline, Codex, and Claude Code work out of the box. Plain chat and tool-driven agents bill at the same rate.
How does billing work?
Usage-based, per token — no tiers, seats, or minimums. You pay a flat 0.75× of our list rate per million tokens (25% off), on both input and output, for everything. The rate is shown on each API key in the console, so what you see is what you're billed. Top up your balance and draw down as you go.
Do you offer deeper discounts at volume?
Yes. High-volume and enterprise customers can get a lower multiplier on their account — the rate is applied the same way and shown on your keys. Contact us to discuss volume pricing.
Which models are available?
Claude Opus and Claude Sonnet, plus a smart model that automatically selects the best model for each request. See the models page for the full list.
Can I run it in my own environment?
Yes. Private deployment delivers Drawbridge as licensed software you run inside your own infrastructure, exposing the same universal API with all supported formats. Contact us for license terms and volume pricing.