Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@mcp-http GET https://api.github.com/users/octocat"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP server exposing a single generic HTTP client tool (http_request) so an AI
agent can make arbitrary HTTP requests from its own environment.
Installable from a git repository via uvx:
uvx --from git+https://github.com/Llamatron2112/mcp-http@main mcp-http-serverOr run locally:
uv run mcp-http-serverPoint your agent at the server over stdio, e.g. in opencode.json or
claude_desktop_config.json:
{
"mcpServers": {
"http-client": {
"command": "uvx",
"args": ["--from", "git+https://github.com/Llamatron2112/mcp-http@main", "mcp-http-server"]
}
}
}Related MCP server: cURL MCP Server
http_requestThe agent controls everything: method and url are required; params,
headers, json_body, data, timeout and follow_redirects are optional.
http_request(method, url, params=None, headers=None, json_body=None,
data=None, timeout=30.0, follow_redirects=True) -> dictReturns {"status", "content_type", "headers", "body", "truncated"}. body is
truncated to HTTP_MAX_BODY (default 1 MB).
A header value of the form ${VAR} is resolved by the server from its own
environment variables. The agent only writes the placeholder, so the real value
never enters the conversation, tool history, or logs.
{"headers": {"Authorization": "Bearer ${API_TOKEN}"}}Example: define the variable on the machine running the server and reference it with any name:
export API_TOKEN="sk-..."The agent keeps full control for security testing: with token, without token (just omit the header), or with a wrong token to test for a 401.
Variable | Default | Meaning |
|
| Maximum response body bytes returned to the agent |
Send an arbitrary HTTP request and return the response.
Single generic HTTP client tool: the agent picks the method, URL, headers, query parameters and body. Returns status, headers, content-type and body.
Supported methods: GET, POST, PUT, PATCH, DELETE, HEAD, OPTIONS, TRACE.
Inputs: url (required), method (required), data, params, +4 more. Returns: object
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
Scored across 1 tool
With only a single tool, there is no possibility of confusion or overlapping purposes. The http_request tool has a clear, distinct role as a generic HTTP client.
The tool name 'http_request' is a clear verb_noun pattern. Since there is only one tool, consistency is trivially maintained, and the name accurately reflects its function.
While a single tool feels thin, it is justified for a focused HTTP client server. The tool is comprehensive and not trivial, but the server offers no additional utilities or convenience wrappers.
The http_request tool supports all common HTTP methods, headers, query parameters, and body payloads, and returns status, headers, and body content. This fully covers the expected domain of an HTTP client with no obvious gaps.
Pay-per-use agent tools: schedule HTTP calls, extract pages, PDFs, OG images, QR codes, short links.
Schedule HTTP requests from an agent: once at an exact time or on cron, with retries and run logs.
Let agents make HTTP requests later: schedule, retry, cron, watch jobs, receive webhooks.
Reliable web access for AI agents: smart HTTP, rotating proxies, and full-browser rendering.
