crewai-stackresolve

CrewAI tools for StackResolve: web intelligence for AI agents.

Choose software for a task, compare vendors, audit agent-readiness, and pull structured company facts. One tool call instead of a search-and-scrape loop.

pip install crewai-stackresolve

Quickstart

from crewai import Agent, Task, Crew
from crewai_stackresolve import stackresolve_tools

analyst = Agent(
    role="Tooling analyst",
    goal="Pick the right software for a task and justify the cost",
    backstory="You evaluate vendors on agent-readiness, not marketing copy.",
    tools=stackresolve_tools(),
)

task = Task(
    description="We need to scrape javascript-heavy sites. Recommend a tool and its cost.",
    expected_output="A recommendation with the AgentReady score and current pricing.",
    agent=analyst,
)

print(Crew(agents=[analyst], tasks=[task]).kickoff())

Tools

Find software for a task, Search the software registry, Compare products, Audit agent readiness, Get company facts, Get company pricing, Find competitors, Research a company.

All of them work without a key, subject to an anonymous rate limit. A free key from stackresolve.dev/developers raises it:

stackresolve_tools(api_key="ar_...")   # or set STACKRESOLVE_API_KEY

Import a single tool if you do not want the whole set:

from crewai_stackresolve import StackResolveAudit
agent = Agent(role="Auditor", tools=[StackResolveAudit()])

Behavior worth knowing

Tools return readable text on failure rather than raising, so a timeout or an exhausted allowance does not end the crew's run. Output is capped at 6,000 characters and stays valid JSON when truncated.

Also available

Hosted MCP server: https://mcp.stackresolve.dev/mcp. LangChain: pip install langchain-stackresolve. LlamaIndex: pip install llama-index-tools-stackresolve. TypeScript: npm install stackresolve.

License

MIT