Octrafic is an open-source CLI for API testing. Point it at any OpenAPI spec or live endpoint, describe what you want to test in plain English, and let it handle the rest - from generating requests to validating responses and exporting a PDF report. No test scripts, no GUI, no mocks. Just a single binary. Works with OpenAI, Claude, Ollama, and any OpenAI-compatible provider
I built Octrafic to make API testing simpler - no test scripts, no GUI, no mocks.
Point it at any API, describe what you want to test in plain English, and the AI agent handles the rest - planning scenarios, running real requests, validating responses, and exporting results.
What it can do:
- Describe tests in plain English - no boilerplate, no config files
- Generate an OpenAPI spec from your source code
- Run in CI/CD pipelines non-interactively with a single command
- Export tests to Postman, curl, or pytest to use in your existing toolchain
- Export PDF reports
- Works with any LLM - OpenAI, Claude, Ollama, llama.cpp. You bring your own key, nothing goes through my servers.
Single binary, no runtime dependencies, fully open-source under MIT.
@juelz I've been testing it with Qwen 2.5 7B and it handles most cases well! For really complex specs a bigger model helps, but for everyday use 7B does the job.
@piroune_balachandran Yeah that's already in there! octrafic scan reads your routes and generates the spec from code. Hook it into CI and it stays in sync on its own. Love that you see it as a source of truth that's exactly where I'm heading with this. Thanks!
Report
@hawierdev Congratulations on the launch! What’s the onboarding experience like for developers new to Octrafic? Can someone start testing an API in minutes?
Report
"No scripts, no GUI, no mocks" resonates — I've lost too many hours writing boilerplate just to validate a single endpoint. The OpenAPI spec auto-generation from source code is a smart move. Does it handle versioned APIs where multiple spec versions coexist?
Report
This is exactly the kind of tool I didn't know I needed. The plain English to test case pipeline is clever — how does it handle auth heavy APIs where half the complexity is just getting a valid token before you can test anything? Also curious how the PDF reports look, that could be genuinely useful for handing off to a client.
Report
Congrats for the launch - this looks super useful and impressive!
Termphin
Hey everyone! 👋
I built Octrafic to make API testing simpler - no test scripts, no GUI, no mocks.
Point it at any API, describe what you want to test in plain English, and the AI agent handles the rest - planning scenarios, running real requests, validating responses, and exporting results.
What it can do:
- Describe tests in plain English - no boilerplate, no config files
- Generate an OpenAPI spec from your source code
- Run in CI/CD pipelines non-interactively with a single command
- Export tests to Postman, curl, or pytest to use in your existing toolchain
- Export PDF reports
- Works with any LLM - OpenAI, Claude, Ollama, llama.cpp. You bring your own key, nothing goes through my servers.
Single binary, no runtime dependencies, fully open-source under MIT.
Give it a shot and let me know what you think.
Termphin
Termphin
@hawierdev Congratulations on the launch! What’s the onboarding experience like for developers new to Octrafic? Can someone start testing an API in minutes?
"No scripts, no GUI, no mocks" resonates — I've lost too many hours writing boilerplate just to validate a single endpoint. The OpenAPI spec auto-generation from source code is a smart move. Does it handle versioned APIs where multiple spec versions coexist?
This is exactly the kind of tool I didn't know I needed. The plain English to test case pipeline is clever — how does it handle auth heavy APIs where half the complexity is just getting a valid token before you can test anything? Also curious how the PDF reports look, that could be genuinely useful for handing off to a client.
Congrats for the launch - this looks super useful and impressive!
Termphin