Private · OpenAI-compatible · pre-launch

The private marketplace for AI inference.

DEX.DO is a token marketplace for AI model access, built on Dark Order infrastructure. Route to the best seller through one OpenAI-compatible API, pay per token as you go, and lock in your future inference cost on the secondary market.

· CODEX-TICK
Order book
Price · SHELL/tickSize · ticks
spread 20 3,500 mid
Secondary: hedge CODEX-TICK → T+30d · settled from this book
At launch · limited capacity

Hedging and semantic orders

Lock the future price of a model on the secondary market, or place semantic orders on AI-event outcomes — settled from the marketplace's own order book. Both open with limited capacity at launch and scale as the book deepens.

Early access

Get on the list before launch

We're onboarding early buyers and sellers ahead of testnet. Tell us about your usage and we'll reach out with access.

No spam. We'll only use this to invite you to the testnet and a demo.

Thanks — you're on the list. We'll be in touch shortly.
FAQ

Straight answers.

An open order book matches you to the best available ask across sellers, instead of charging a fixed markup on every token. And you get something they don't offer at all — the ability to hedge the price.

The buyer-facing marketplace runs on supply whose resale is permitted by provider terms (open models, wholesale, resale-allowed APIs). It's the same category of service as any model router.

DEX.DO inherits Dark Order privacy — shielded prices and order sizes, no front-running — on an anonymous L1 with no account wall. Requests reach the provider directly; the chain only meters usage.

Most are spot-only. DEX.DO adds a secondary market so you can hedge your future cost — plus a private on-chain order book and a rewards program for early participants.

Access to a model is sold in fixed ticks. You pay per tick as the work streams, so there's nothing to trust up front and your downside is capped.

The marketplace goes live on testnet first, with test tokens, ahead of mainnet. Join early access to get in with the first cohort.