TUFTech Docs

Quick start

Three steps to get connected.

1. Sign in to the console, open “Token Management”, then create and copy an API Key.Key shown once
2. Set the client Base URL to https://api.tuftech.org.
3. Pick the path for your client: OpenAI-compatible use /v1/chat/completions or /v1/responses, Claude-compatible use /v1/messages.

curl https://api.tuftech.org/v1/chat/completions \
  -H "Authorization: Bearer $TUFTECH_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gpt-5.5",
    "messages": [{"role": "user", "content": "Hello"}]
  }'

Endpoints

All traffic goes through https://api.tuftech.org; choose the path that matches your client.

Common paths: /v1/chat/completions · /v1/responses · /v1/messages · /v1/embeddings · /v1/images/generations · /v1/videos

Groups & ratios

Actual charge = model base price × the group ratio of your token. The table below lists user-selectable groups, sorted by ratio ascending; the rate-limit column is each group's “requests / successes” cap per 1 minute(s).

Lower ratio is cheaper. Ratios are taken from the console GroupRatio config; the usage log is authoritative for billing.

Rate limits

Rate limiting is counted per user account; multiple API Keys under one user share the quota, and different users don't affect each other.

• Platform default: up to 120 requests and 100 successful requests per 1 minute(s).
• Each group can override the default — see the rate-limit column in “Groups & ratios” above (e.g. veo is clearly lower, to control high-cost tasks).
• Hitting the limit returns 429; just retry shortly.

Veo video models

The Veo group includes veo_3_1, veo_3_1-fast-fl, veo_3_1-fast-fl-hd, billed per call.

After submitting a task you get a task_id; poll status with /v1/videos/{task_id}. Failed or cancelled tasks are refunded per platform logic.

Troubleshooting

On a failed call, check the usage log first: model name, group, status code, error message, request time, charge record.

If you can't resolve it, contact an admin with the log screenshot or key details.

FAQ

9 entries.