Running Recurse as a service, AI Agents

Recurse Agency Implementation, Building AI Agent Services

Learn how to architect, test, and deploy custom AI agents using Recurse's declarative agent.yaml manifest for your clients. This course teaches agencies how to build productized agent services, manage candidate testing workflows, and structure pricing around per-minute execution costs to maximize margins on automation projects.

Open the decision record for Recurse

What does running Recurse for clients commit you to?

Published figures for this service. Blank fields are not published.

Monthly tool cost
Vendor cost basis: $5 of included runs on every new account, then usage-based per-request billing for model calls and secure execution. No subscription or idle charges are published. Client pricing and markup are Not modeled, the agency must supply its own billing rate.
Time to first value
Not published
Payback
Not modeled
Guided implementation
8 hours

Is Recurse worth running as a client service?

The evidence supports an engineering-led managed service built on Recurse's declarative agent.yaml contract, candidate scoring loop, and deployment as tools, MCPs, or bots, with a $5 starting run credit and per-request billing as the vendor cost. What remains unknown is any agency-side pricing, labor cost, margin, volume, or payback timeline, none of which the supplied data publishes.

An agency-fit judgement for reselling this service. It is separate from the tool description on the decision record.

Before you start

What has to be in place before the first client engagement.

Tools and subscriptions

  • Recurse account funded with the published $5 of starting runs
  • Codex integration for the coding agent design loop
  • Claude integration listed among Recurse's supported integrations
  • Python source file (tools.py) for registering custom agent tools
  • Recurse SDK or API access for deploying agents as tools, MCPs, or bots

People and inputs

  • Engineer able to author declarative agent.yaml manifests with pinned input/output schemas
  • Prompt author for prompt.md referenced by the manifest
  • Code review process for custom Python tool registration in tools.py
  • Validation harness able to score candidate outputs against a measurable bar

Included with the course

7 working documents for delivering this service.

  • Agent Project Scoping Worksheetworksheet
  • agent.yaml Manifest Template Librarytemplate
  • Candidate Testing and Quality Bar Checklistchecklist
  • Recurse Deployment SOP (Tools, MCP, Bots)sop
  • Usage-Based Pricing Calculator for Client Proposalsworksheet
  • Custom Python Tools Registration Guideguide
  • Agent Monitoring and Iteration Playbookguide

Listed by name. These documents are not yet published as individual downloads.