AI ASSISTANT

Lite

AI that speaks your API's language.

Generate requests from plain English, explain response payloads, write pre-request scripts, and debug errors — all without leaving your workflow.

Rostyman AI assistant explaining an HTTP response inside the right panel

What you can do with the AI assistant

Generate requests from plain English

Describe what you want to test — "create a POST request to register a new user with email, name, and password fields" — and Rostyman builds the full request including URL, method, body schema, and headers.

Example: "Get all products in the Electronics category with price below $100, sorted by rating"

Explain response payloads

Paste any response JSON or select it in the response viewer. Ask "what does this response mean?" and get a plain-English explanation of every field, including inferred types and relationships.

Example: "What does the nested meta.pagination object mean in this response?"

Write pre-request scripts

Describe the script you need — "compute an HMAC-SHA256 signature for this request" or "refresh the OAuth token if it expires in less than 5 minutes" — and get working rm.* API JavaScript you can run directly.

Example: "Set a dynamic timestamp header for AWS request signing"

Debug API errors

Paste an error response, a 500 trace, or an unexpected result. The AI explains what went wrong, what the API server likely returned and why, and suggests how to fix the request.

Example: "Why am I getting 401 even though I set the Authorization header?"

Generate test assertions

After receiving a response, ask the AI to write post-response test scripts using the rm.test() API — asserting status, checking specific fields, validating schema structure.

Example: "Write tests to validate this user response has all required fields"

Convert cURL to request

Paste any cURL command and Rostyman's AI will parse it into a complete request — method, URL, headers, body, and auth — ready to send and save.

Example: "Convert this curl -X POST command to a request"

AI tool-use — chat that builds

The assistant doesn't just answer questions about your workspace — it can edit it. Ask for a collection, a folder, a request, or a full visual workflow and it appears in your sidebar. Every action is auditable in the chat transcript.

create_collection

Spin up a new collection with name + description from a chat prompt — "make a collection for the Stripe API".

create_folder

Add a folder under any collection with a single message — useful for grouping by feature or version.

create_request

Generate a fully-formed HTTP request — method, URL, headers, body, auth — directly into the folder you point at.

create_workflow

Build a visual workflow from a description — "OAuth login → fetch profile → call /orders". Nodes, edges, and assertions all wired.

get_collection

List the folders and requests in any collection so the assistant can reason about what already exists before adding more.

Audit-friendly

Every tool call is logged in the chat transcript with arguments and result — undo, copy, or replay any step.

AI for databases

The same assistant is wired into the database client. It sees your live schema and uses it to write SQL that actually runs.

Plain-English → SQL or MongoDB

Ask a question in plain English; get a runnable query in your dialect — Postgres, MySQL, MSSQL, SQLite, CockroachDB, or MongoDB. The schema is the source of truth.

Schema analysis

One click audits your schema for missing indexes, missing constraints, and type mismatches with per-table recommendations.

Stored procs, functions, triggers, views

Describe behavior in English; get a CREATE statement in the right dialect, ready to run.

Generate full CRUD API collections

Pick the tables; get a complete REST collection (List / Get / Create / Update / Delete) wired to {{baseUrl}}, dropped into your sidebar.

See the full database client →

5 AI providers — all BYO key

Your API key, your data. We never proxy your requests through Rostyman servers.

OpenAI

GPT-4o, GPT-4o Mini, GPT-4.1, GPT-4.1 Mini, o3, o4-mini

Connect with your own API key. Supports the latest models with function calling.

Requests go to OpenAI servers. Rostyman never stores your API key in plaintext.

Anthropic

Claude Opus 4.7, Claude Sonnet 4.6, Claude Haiku 4.5

Use Anthropic's Claude models. Excellent at understanding complex API structures and writing scripts.

Requests go to Anthropic servers. Rostyman never stores your API key in plaintext.

Google Gemini

Gemini 2.5 Pro, Gemini 2.5 Flash, Gemini 2.0 Flash

Use Google's Gemini models with a Gemini API key from Google AI Studio.

Requests go to Google servers. Rostyman never stores your API key in plaintext.

GitHub Copilot

GPT-4o, GPT-4.1, Claude Sonnet 4.6 via Copilot

Use your existing GitHub Copilot subscription — no separate API key needed.

Requests go to GitHub Copilot servers. Rostyman never stores your API key in plaintext.

Maximum Privacy

Ollama (Local)

Any Ollama model — dynamically fetched from localhost:11434

Run AI entirely on your machine. Zero network requests. Perfect for air-gapped environments or maximum privacy.

Completely private. Nothing leaves your machine. Requires Ollama installed locally.

Already have a CLI AI tool?

Rostyman can drive Claude Code, GitHub Copilot CLI, Gemini CLI, or OpenAI Codex if they're already installed on your machine — no second API key required.

Claude Code

GitHub Copilot CLI

Gemini CLI

OpenAI Codex

Privacy guarantee

Rostyman never sends your request data or collection contents to our servers — ever.
AI requests go directly from your machine to your chosen provider (OpenAI / Anthropic / Google / Copilot).
With Ollama or a local CLI tool, everything runs locally — not even the AI provider sees your data.
Your API keys are stored encrypted in your local SQLite database.
We only collect anonymous install metrics (version, DAU) — never request or response content.

Start using AI-assisted API testing today

Bring your own API key or use a local CLI tool — included free in Rostyman Lite.

Download Free