Advanced AI solutions for modern business.
We design, build and run AI agents that reason before they act — and that you can check when they do.
class Agent: def act(self, goal): plan = self.reason(goal) evidence = self.retrieve(plan) return self.decide(plan, evidence)
What we do
Learn more about our AI solutionsRational Agents is an AI thought works company.
Agentic products & solutions
We design, build and run agentic apps that reason before they act. We orchestrate them with LangGraph agents, for workflows and data pipelines that are more reliable than LLMs alone.
- Signal Flare
Finds the roles, reaches the hiring managers, tracks every reply.
- Coleslaw
Cards in; reviewed, tested code out.
- Muffin
Chunking that fits every document — coming soon.
Research
We research AI systems that are both powerful and explainable, integrating neural networks with symbolic logic.
Read our researchForward deployed engineering (FDE)
Our engineers work alongside your team to deliver AI systems that solve real-world problems and drive measurable results.
- AI integration
- Seamlessly integrate state-of-the-art AI solutions into your existing business workflows.
- Data analytics
- Transform your raw data into actionable insights with our advanced analytics solutions.
- Custom solutions
- Tailor-made AI applications designed specifically for your business needs.
Claude Code plugins
adr-consensus
Drive an Architecture Decision Record to a READY, NEEDS_DECISION or NOT_READY verdict with a quorum of reviewing models.
replicants
Engineering agents and skills — implementer, analyst, QA, architect, deploy — that bind to whatever repository they run in.
claude plugin marketplace add rational-agents-ai/rational_agents_plugins --scope projectclaude plugin install adr-consensus@rational-agents --scope projectclaude plugin install replicants@rational-agents --scope projectRun from the repository you want the plugins in. Recorded in its .claude/settings.json, so everyone who works there gets the same setup.