Help center
Answers to common questions about getting a key, billing, usage limits, and errors. For full endpoint details, see the API reference.
Getting started
How do I get an API key?
POST /v1/keys/free mints a free key — 100 calls/day, no card required. See the free-key endpoint for the request shape.
How do I upgrade to Pro?
Pick a plan on the pricing page — $49/mo, monthly, quarterly, or annual billing via Paddle. Pro keys get 5,000 calls/month.
Billing & subscription
How do I cancel my subscription?
Use the Manage subscription link in your Paddle receipt email. Cancelling stops future billing; your key runs until the paid period ends — no lock-in.
Can I switch billing cadence (monthly / quarterly / annual)?
Email contact@savetoken.org to switch cadence. Your key runs until the current paid period ends.
Do you charge for my LLM tokens?
No. SaveToken returns a smaller version of your text; you send that to Claude, OpenAI, or whichever provider you use, with your own key. We never see or bill for that call.
Can I get a refund?
See the refund policy — contact contact@savetoken.org within 7 days of a charge.
Usage & limits
| Limit | Applies to |
|---|---|
| 300 requests / minute / key | Every key — a burst guard, not a plan limit. |
| 100 requests / day / key | Free-tier keys. |
| 5,000 requests / month / key | Pro keys, rolling 30-day window. |
| 20 mints / day / IP address | POST /v1/keys/free. |
What happens if I hit my plan's call limit?
Extra calls are declined until usage drops back under the limit — no overage charge. If a Pro key hits the cap, email contact@savetoken.org and we'll bump it.
What counts as one call?
One POST /v1/optimize request. How much it reduces depends on the input — structured logs, JSON, and tool schemas shrink the most.
Errors
Every error response has the same shape — a code field and a request_id to reference if you contact support.
| Status | code | What it means | Fix |
|---|---|---|---|
| 400 | BAD_REQUEST | The request body is malformed or violates a field constraint. | Check the message field and the request schema. |
| 401 | UNAUTHORIZED | The API key is missing, invalid, or revoked. | Confirm the Authorization: Bearer header and that the key hasn't been revoked. |
| 429 | RATE_LIMIT | A rate or quota limit was reached. | See usage & limits above — the message names which limit was hit. |
| 500 | INTERNAL | A server-side failure. | Retry the request; if it persists, email contact@savetoken.org with the request_id. |
Privacy & data
Does SaveToken store my data?
No. SaveToken logs token counts, dollars saved, and the verdict per call — never the request or response text itself. See the privacy policy for full detail.
Do you see my LLM provider's API key?
No. SaveToken processes the text you send to produce the optimized version; it never sees or stores your downstream LLM provider key.
Does it work with non-Claude providers?
Yes. SaveToken never calls the model itself, so the optimized text works with Claude, ChatGPT, Gemini, or anything else you send it to.
Does it break prompt caching?
No. The stable prefix (system prompt, tool definitions) is marked by you and never touched — only the dynamic tail is compressed, so cached reads stay cheap.
Contact
request_id (if you have one) and we'll get back to you.Building against the API? An agent can file an issue directly — no email client needed: POST /v1/support.