Welcome to the Wokey Blog

Wokey Team · 2026-06-13 · Updated 2026-09-25

One API key for the official Claude, GPT and other models, below list price, with responses you can check yourself came from the official upstream. That is why Wokey exists. This is the first post on the blog: a quick look at what we're building, why, and what you can expect here next.

The problem we're solving

When people shop for a cheaper API, price is rarely the real worry. Two other things are:

  • Is it the official model? A relay can put anything in the model field, and answer quality is a matter of opinion.
  • Was the output changed? Truncated, swapped, or mixed with another model's output.

Here is how Wokey handles that:

  • Official upstream. Requests go to the model vendor's official API, for example api.anthropic.com for Claude.
  • Verifiable. On the raw passthrough path, the official exit signs the upstream response bytes inside an AWS Nitro Enclave, and streaming responses end with event: tee.proof. The open-source verifier checks it on your own machine, with no trust in us required.
  • One key, pay per use. OpenAI- and Anthropic-compatible. Every model page lists its price, and every call's billing detail is on record.

In one line: don't trust, verify.

One-minute quickstart

Once you have a key, calling Wokey looks exactly like the OpenAI API you already know:

curl https://api.wokey.ai/v1/chat/completions \
  -H "Authorization: Bearer $WOKEY_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "claude-sonnet-5",
    "messages": [{ "role": "user", "content": "Hello, Wokey!" }]
  }'

Claude Code users set ANTHROPIC_BASE_URL to https://api.wokey.ai (no /v1). See Claude Code setup.

What's coming next

  • Setup and troubleshooting for Claude Code, Codex and other clients
  • The logic behind routing and pricing
  • How response proofs work and how to verify them yourself

Thanks for reading, see you in the next post.