Rails AI Agents
Production patterns for building AI agents in Ruby and Rails. These guides focus on provider boundaries, tool calling, background execution, human approval, observability, and cost control.
MODEL BOUNDARY
Anthropic SDK, Gemini HTTP clients, parser tests, and provider-specific execution shapes.
TOOLS
Function calling, authorization, idempotency, write approvals, and audit-friendly tool registries.
OPERATIONS
Solid Queue isolation, resumable runs, step traces, cost attribution, and deploy-safe execution.
Building AI Agents in Ruby with the Anthropic SDK
The full production shape: agent loop, tool classes, streaming, MCP, authorization, approval gating, and tests.
Gemini API in Ruby Without an SDK
Faraday transport, Interactions API state, function calls, background execution, streaming, and guardrails.
Running AI Agents with Solid Queue
Queue isolation, worker throttling, resumable runs, human-in-the-loop pauses, and per-account cost tracking.
Recurring Agent Work
How Solid Queue recurring jobs affect polling agents, scheduled inference, and at-least-once execution safety.
API Integrations for Agent Tools
Agents become useful when their tools can safely read and write CRM, ERP, accounting, and internal data.
Building an AI workflow inside Rails?
I help teams turn demos into production agent workflows with safe tools, durable state, background execution, and measurable cost controls.
View AI agent services