AgentBodega MCP
OfficialREADME.md
# AgentBodega MCP
[](https://smithery.ai/servers/agentbodega/agentbodega)
Give agents a live menu of paid, x402-ready tools from
[AgentBodega](https://agentbodega.store).
This MCP server lets agents discover AgentBodega endpoints, compare prices,
plan the cheapest useful route for a task, inspect required inputs and response
formats, and generate ready-to-run x402 HTTP calls. Agents get the buying
instructions for each resource without guessing schemas or scraping docs.
This public repository contains only the AgentBodega MCP package. The hosted
AgentBodega API runs separately at [agentbodega.store](https://agentbodega.store).
## What Agents Can Find
- Social media lookups and research tools.
- Real-estate listing search.
- Service status and cloud status checks.
- Domain, launch-readiness, and agent-discoverability audits.
- x402 inspection, directory packaging, and hosted artifact utilities.
The MCP server does not execute paid calls. It returns the endpoint contract,
example request, payment requirements, and snippets an agent can run against the
hosted API. L402 is exposed as an additional discovery and payment-handoff
surface; agents should treat it as discovery-only unless the live L402 document
reports a configured issuer.
## Install
```bash
npx @agentbodega/mcp
```
Remote MCP clients that support Streamable HTTP can connect directly to:
```text
https://agentbodega.store/mcp
```
SSE endpoint:
```text
https://agentbodega.store/mcp/sse
```
Optional environment:
```bash
AGENTBODEGA_BASE_URL=https://agentbodega.store
```
Standalone agent skill:
```bash
curl -sS https://raw.githubusercontent.com/agentbodegastore/agentbodega/main/agent-skills/agentbodega/SKILL.md
```
## Network access
This package intentionally uses outbound HTTPS requests at runtime to read
AgentBodega's live catalog, x402 discovery document, L402 discovery document,
and OpenAPI contract. It does not run install scripts, spawn shells, bundle
dependencies, or execute paid calls. By default it connects to
`https://agentbodega.store`.
## Claude Desktop
```json
{
"mcpServers": {
"agentbodega": {
"command": "npx",
"args": ["-y", "@agentbodega/mcp"]
}
}
}
```
## Tools
- `agentbodega_plan_task` - use first when an agent has a plain-language task
and wants AgentBodega to choose the cheapest useful route. It calls the free
`/api/plan` planner and returns a recommended endpoint, alternatives,
required inputs, examples, and x402 handoff notes. It never executes paid
calls.
- `agentbodega_search_catalog` - use first to find the right paid endpoint. It
searches the live directory by text, department/category/tag, and maximum
USD price, then returns endpoint summaries with required inputs and example
requests. It never executes paid calls.
- `agentbodega_get_endpoint` - use after search when you know an endpoint key,
path, title, or full URL. It returns the full contract: method, URL, price,
required inputs, schemas, example request, output example, and long
description. It does not call the endpoint.
- `agentbodega_call_snippet` - use after selecting an endpoint to generate
ready-to-edit curl and JavaScript requests with an `X-PAYMENT` placeholder
and x402 verification notes. It does not fetch a challenge, pay, or execute.
- `agentbodega_payment_guide` - use when an agent needs the AgentBodega payment
model before making paid HTTP calls. It explains x402 discovery, L402 handoff
status, settlement, blockchain-only balance policy, and optional challenge
curl.
## Resources
- `agentbodega://catalog` - compact live catalog.
- `agentbodega://x402` - live x402 discovery document.
- `agentbodega://l402` - live L402 discovery document.
- `agentbodega://openapi` - live OpenAPI document.
## Development
```bash
npm install
npm test
```
TDQS
A4.7/5.0
Scored across 4 tools
Disambiguation5/5
Each tool has a clearly distinct purpose in the workflow: search catalog, get endpoint details, payment guidance, and code snippet generation. No overlap or ambiguity.
Naming Consistency5/5
All tools follow a consistent 'agentbodega_verb_noun' pattern using snake_case, making them predictable and easy to understand.
Tool Count5/5
With 4 tools, the set is well-scoped for the server's goal of guiding users through paid endpoint discovery and preparation. No tool feels extraneous or missing.
Completeness5/5
The tools cover the entire workflow from search to snippet generation, along with payment guidance. No obvious gaps given the server's stated purpose of not executing paid calls.
Maintenance
ActivityStale
ResponsivenessUnresponsive