Documentation & Quickstart

Getting started with Mimicly

This is a conceptual guide: Mimicly isn't built yet, so the endpoint paths and response shapes below show the idea, not a live API. About the concept.

1. Point your test environment at Mimicly

Swap the base URL your AI client uses in development, staging and CI. Your production configuration stays untouched.

client.ts

const client = new OpenAI({
  apiKey: process.env.OPENAI_API_KEY,
  baseURL: process.env.AI_BASE_URL,
});

2. Record a real interaction

Run your application against the real provider through the Mimicly recorder. The request, response, model, endpoint, parameters and any tool calls are captured for that interaction.

3. Convert the recording into a deterministic mock

Choose the matching strategy that fits each test case:

Exact match — identical payload returns the recorded response.
Endpoint match — matches HTTP method and path, regardless of body.
Pattern match — matches when the prompt text contains a phrase you define.
Scenario match — your test triggers a named scenario explicitly.

4. Simulate failures & edge cases on demand

Configure a scenario to return a 429 rate limit, a 500 error, a timeout, an invalid-key response, or a malformed response, without waiting for the real provider to misbehave.

5. Fast, repeatable CI/CD pipelines

Set the same base URL in your GitHub Actions, GitLab CI, or local Docker tests. Mock responses return immediately, with zero token spend.