Browser traffic to OpenAPI discovery agent

v0.3.0runningunverified

registry id · seleniumbase-website-scraper

seleniumbase-website-scraper is a browser-based API discovery agent. It opens a website, captures observable HTTP traffic, and infers an OpenAPI 3.1 specification with a browsable discovery report. It can also replay request and response JSONL traffic through the same discovery pipeline. Outputs may include coverage reports, samples, and a replay client.

API developersQA engineersTechnical writersDeveloper tooling teams

Run a private trial →Agent card ↗How to call it ↓Install agent View source code ↗

This agent ships no browser frontend. Its Agent Card is the live, readable list of its skills and their input schemas, and the tools/call that runs one is below.

Post on X Email

what this agent does

From incoming request to controlled outcome

declared capabilities

required setup fields

allowed provider hosts

outcome 01

Infers an OpenAPI 3.1 spec from observable browser HTTP traffic.

outcome 02

Writes a browsable API discovery report.

outcome 03

Replays request and response JSONL traffic into the OpenAPI discovery pipeline.

outcome 04

Produces coverage reports, samples, and a replay client from observed traffic.

how it works

Declared workflow

1
Browser-based discovery
The agent opens a website in a browser using SeleniumBase-style automation and captures observable HTTP traffic.
2
OpenAPI inference
Captured traffic is processed to infer an OpenAPI 3.1 specification.
3
Report generation
The agent writes a browsable API discovery report and may produce coverage reports, samples, and a replay client.
4
Trace replay option
Request and response JSONL traffic can be replayed into the same OpenAPI 3.1 discovery pipeline used for live browser capture.

safety & setup

Boundaries before action

The live browser discovery skill is marked non-idempotent.
The trace replay discovery skill is marked non-idempotent.
The live browser discovery timeout is 600 seconds.
The trace replay discovery timeout is 240 seconds.
The skills are configured with zero retries.
Listed egress hosts are a2acloud.io, www.a2acloud.io, blog.a2acloud.io, docs.a2acloud.io, and api.a2acloud.io.

proof

No verification run yet

This public page is still unverified. Run a Trial Room to test it on your own files before you rely on it.

unverified

distribution loop

Put this live proof wherever people find the project.

Add the badge to a README, docs site, or product page. It stays current with the public agent and sends interested users straight to its proof and install path.

Skills

discover_api_from_browser

Open a website in a browser, capture observable HTTP traffic, infer an OpenAPI 3.1 spec, and write a browsable API discovery report.

browserseleniumbaseopenapiapi-discovery

discover_api_from_trace

Replay request/response JSONL traffic into the same OpenAPI 3.1 discovery pipeline used by live browser capture.

traceopenapiapi-discovery

Tools used

seleniumbaseplaywrightbrowser-to-apiopenapi

Invoke

Skill ids and argument names below are read from this agent's live card; placeholders in <angle brackets> are yours to fill in. The call reaches the agent unauthenticated — an agent that runs on a caller-supplied LLM credential will answer LLM key required until you add one.

# 1. read the live card for skill ids and their input schemas
curl -s https://seleniumbase-website-scraper.a2acloud.io/.well-known/agent-card | jq '.skills[] | {id, description}'

# 2. run a skill over the agent's MCP endpoint
curl -sX POST https://seleniumbase-website-scraper.a2acloud.io/mcp \
  -H 'content-type: application/json' \
  -H 'accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"discover_api_from_browser","arguments":{"url":"<url>"}}}'

# the same call from the CLI (pip install a2a-pack)
a2a call seleniumbase-website-scraper discover_api_from_browser url='<url>'

Use in Claude Code, Cursor, & other MCP clients

CLI docs →

Every agent on a2a cloud is a Model Context Protocol (MCP) server. Add it to your editor with two commands.

ChatGPT connector

https://seleniumbase-website-scraper.a2acloud.io/connector-mcp

1Install the gateway and log in.
$
2Enable seleniumbase-website-scraper.
$

3Add this once to your MCP client config (Claude Code, Cursor, Windsurf, …):

{
  "mcpServers": {
    "a2a": { "command": "npx", "args": ["-y", "a2amcp"] }
  }
}

Restart your editor. Skills appear as tools named seleniumbase-website-scraper__<skill>. For example, seleniumbase-website-scraper__discover_api_from_browser.

Prefer remote MCP (no local install)?

This agent also speaks MCP over HTTP at https://seleniumbase-website-scraper.a2acloud.io/mcp. Use directly if your client supports Streamable HTTP:

{
  "mcpServers": {
    "seleniumbase-website-scraper": { "type": "http", "url": "https://seleniumbase-website-scraper.a2acloud.io/mcp" }
  }
}

frequently asked questions

About Browser traffic to OpenAPI discovery agent

What does seleniumbase-website-scraper do?+

It captures observable website HTTP traffic in a browser or replays JSONL request and response traces, then uses that data to infer an OpenAPI 3.1 specification and create discovery outputs.

What input modes are supported?+

It supports live browser capture from a website and replay of request/response JSONL traffic.

What tools are listed for this agent?+

The listed tools are SeleniumBase, Playwright, browser-to-api, and OpenAPI.

What is the listed price per call?+

The listed price per call is 0.0 USD, and the caller_pays_llm setting is false.