Built to scale with your team across every stage

Pro accounts can either auto-purchase additional credit packs (off by default) or fall back to soft-blocking with a clear in-app banner. Your existing Code Protection activity and audit logs are never deleted. Free accounts soft-block once the cap is reached and can upgrade in one click.

Credit packs are pay-as-you-go top-ups for any plan. They never expire while your subscription is active and consume after your 7-day Code Protection allowance is exhausted. Common pack sizes: 25,000 Code Protection uses ($25), 100,000 Code Protection uses ($89), 500,000 Code Protection uses ($399).

Subscriptions renew automatically each billing cycle. If you cancel, your seats remain active until the end of the current period and downgrade cleanly to Free. There is no data loss on downgrade.

Yes. Pretense is an OpenAI-compatible local proxy. Set ANTHROPIC_BASE_URL, OPENAI_BASE_URL, or the equivalent variable for your tool, and Pretense intercepts every request. Confirmed compatible with Claude Code, Cursor, GitHub Copilot, Continue.dev, and any custom client that respects standard env vars.

Yes. Pro starts with a 14-day free trial, no credit card required. You can downgrade back to Free at any time, and your audit history is retained for 90 days regardless of plan changes.

You will keep Pro features for the rest of the current period and downgrade automatically at the start of the next cycle. You will not be charged for the downgraded month, and any unused token allowance from your previous Pro tier is simply not carried over.

Nothing leaves your machine until after Code Protection runs. The proxy runs locally. Your code never touches our servers.

Yes. Pretense is a local OpenAI-compatible proxy. Any tool with a configurable API base URL works: Claude Code, Cursor, Copilot, custom scripts.

Yes. 14-day full Enterprise trial, no credit card required. Contact jimmy@pretense.ai.

No. Code Protection is identifier-only using a deterministic hash. The LLM sees structurally identical code -- only variable/function/class names differ. Response quality is unchanged.

Yes. The CLI scanner and Code Protection engine have no usage limits for local use. Only the cloud proxy has a 1,000 Code Protection uses / 7 days limit on the free tier.

Nothing. Pretense runs on your machine. Code Protection runs locally before any network request. We never see your code.

Enterprise plan includes on-prem deployment with a private Docker image. Pro is cloud-assisted only.

Private AI deployment costs $50K+/month for ChatGPT-equivalent quality. Pretense protects your code going to ANY provider for $29/seat.

No. LLMs process syntax and structure, not identifier semantics. Your function _fn4a2b() gets the same quality suggestions as getUserToken().

Pro includes 100,000 proxy Code Protection uses per 7-day rolling period. If you exceed that limit, additional uses are billed at $0.001 each (one-tenth of a cent). There are no surprise charges -- you will see usage warnings at 50%, 75%, and 90% before any overage applies.

Yes. The Starter free tier runs entirely on your machine with no usage limits for local use. Enterprise includes first-class support for Docker and Kubernetes on-prem deployment.

Sub-millisecond Code Protection overhead. The proxy adds fewer than 5ms latency on a local machine. Most users report no perceptible difference.

Any OpenAI-compatible API: Claude (Anthropic), ChatGPT (OpenAI), Gemini (Google), Mistral, local Ollama. One proxy, all providers.

Yes, and only by you. The Code Protection map is stored locally (or in your self-hosted store). No third party can reverse your protected identifiers.