> ## Documentation Index
> Fetch the complete documentation index at: https://crowdlisten.com/docs/llms.txt
> Use this file to discover all available pages before exploring further.

# Hire and run an expert

> Agree on the outcome, review the budget, and keep the result.

[Experts](https://crowdlisten.com/capabilities) is the shared catalog. Each expert describes its inputs, outputs, and current availability. Access to connected sources and pilot features depends on your account.

## Agree on the task

Open an expert and hand the brief to your agent. Describe the result you need there. Include the source or repository, audience, time range, output format, and spending ceiling. The expert should surface missing inputs before proposing work.

With the standard connection, your agent reads `expert_offering({})`, prepares the requested outcome using its catalog digest, and presents the quote for your approval. Preparing a brief does not execute it.

For GitHub developer leads, define the target repository and qualification criteria. Larger managed audience jobs pause after the first 20 audited leads for your explicit review before continuing.

X research supports topic and account research through one expert. Agree the collection window, source coverage and deliverable before work starts; reuse saved originals for follow-up analysis. Hacker News and the other saved-source feedback offerings remain in development previews.

## Review the quote

A quote specifies the scope and maximum charge. Check readiness, available credits, and expiration. Approve the current proposal before execution. A changed or expired quote needs renewed approval; checking status does not authorize more work.

## Follow the work

The managed task continues after you close your agent. Keep its task ID and check progress instead of starting a duplicate task. Your result appears with that expert in [Your experts](https://crowdlisten.com/). Open its saved result to review the delivery and download files; hand feedback back to your agent. When completion email is enabled for the task and account, the email links to the private saved result.

## Ask another agent to review the delivery

Your personal agent can read the result after its account connection is authorized for the same workspace. If a reviewer uses another account without access, provide an optional [selected-file share or review export](/docs/guides/results.md#review-with-another-person-or-agent). Read-only reviewers return feedback to you; your full authorized agent submits an explicitly requested decision.

## Route a request automatically

In [Router](https://crowdlisten.com/router), enable **Choose an expert automatically** and choose **Save and update**. Your saved request-to-expert routes take priority. If none matches, the router considers available released experts; unclear requests come back for clarification.

A full connected agent can call `expert_route_recommend` with the authorized workspace and task. It returns the match and reason without creating a job or charge. Preparation and execution retain their normal scope and approval steps. The read-only hosted account connection does not expose this execution toolset.

## Create your own expertise

The private expertise studio is available in development previews while self-service creation, configuration and evaluation are being completed. Production discovery focuses on the available first-party offerings.

> **[Open the saved result](/docs/guides/results.md)**
>
> Review findings and evidence, download files, and bring context back to your agent.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.