Appearance
A.R.C. In One Look
A.R.C. (Agentic Reasoning Core) is the open-source, distributed intelligence orchestration engine that lets you transform weeks of infrastructure setup into a single command. It is built for teams that want a production-ready "Platform-in-a-Box" without surrendering control of their stack—and who expect every service, codename, and circuit breaker to ship pre-wired.
- Tagline: An open-source, modular, Platform-in-a-Box for distributed intelligence.
- Focus: Deliver minutes-to-production workflows, full customizability, and an entirely open-source foundation with chaos-tested resilience.
- Quick Wins: ⚡ Minutes to deploy · 🔧 Fully customizable · 🌐 Open source · 🛡️ Chaos-ready
🧠 What is A.R.C.?
A.R.C. is a distributed intelligence orchestration engine designed to remove infrastructure friction so you can focus on what your agents think and do. With a single CLI-driven experience you get a cohesive, battle-tested ecosystem that ships all essential services pre-wired and production-hardened. The platform maps business-critical roles to named services so you always know who is guarding the door, minding the logs, or keeping the neural net honest.
- 🏗️ Platform, Not Library: Compose a complete ecosystem with one command—no manual wiring of services.
- ⚡ Minutes, Not Months: Go from idea to deployment immediately while A.R.C. handles the plumbing.
- 🔧 Your Stack, Your Rules: Everything is standards-based and swappable with zero vendor lock-in.
- 🧭 Role-Aware Services: Heimdall (Traefik) gates ingress, Sherlock (LangGraph) reasons, T-800 (Chaos Mesh) pressure-tests the cluster, and Friday (SigNoz) shows the vitals.
✨ Why Choose A.R.C.?
- Truly Open-Source: 100% FOSS (Apache 2.0, MIT, MPL). No BSL, "source-available", or non-compete traps.
- Platform-in-a-Box: Heimdall (Traefik), J.A.R.V.I.S. (Kratos), Nick Fury (Infisical), Mystique (Unleash), and more arrive pre-configured.
- Modular & Pluggable: Extend and swap components while staying on open standards like OpenFeature and OpenTelemetry.
- Polyglot Architecture: High-performance Go gateway and CLI paired with the Python AI ecosystem for agent logic.
- Cloud-Native: Containerized, horizontally scalable, and deployable everywhere through a single
arc run. - Event-Driven: Durable streaming via Dr. Strange (Pulsar) plus real-time messaging with The Flash (NATS).
- Distributed Intelligence: Build and orchestrate multi-agent systems with Sherlock, Scarlett, Alfred, and Gordon Ramsay working together.
- Stateful Agents: Persist context with Oracle (Postgres + pgvector) and Sonic (Redis) for semantic recall and hot state.
- Self-Hosted & Resilient: Own your data, keep everything on-prem if needed, and let the T-800 (Chaos Mesh) validate uptime.
🛠️ The A.R.C. Journey
- Design Your Agent — Define capabilities, context windows, and reasoning patterns declaratively. (Focus: capabilities & reasoning)
- Orchestrate Services — Let A.R.C. auto-compose Heimdall, J.A.R.V.I.S., Mystique, and friends into a cohesive stack. (Benefit: pre-configured infrastructure)
- Implement Intelligence — Ship unique agent logic while Sherlock, RoboCop, and Gordon Ramsay handle reasoning, guardrails, and critique. (Approach: logic over plumbing)
- Deploy & Scale — Launch with built-in monitoring, tracing, chaos drills, and scaling; no DevOps fire drills. (Result: production-ready instantly)
🔍 Traditional vs. The A.R.C. Way
| Traditional Build | The A.R.C. Way |
|---|---|
| ⏰ Days spent wiring infrastructure | ⚡ Single-command deployment |
| 🔧 Manual service integration | 🔗 Heimdall + friends pre-integrated |
| 🔒 Risk of vendor lock-in | 🆓 Open-source freedom |
| 🐛 Painful debugging cycles | 🔍 Sherlock + Friday trace issues |
| 📈 Scaling bottlenecks | 🚀 Auto-scaling infrastructure |
| 👓 Limited observability | 📊 OTEL + SigNoz built in |
| 💸 High maintenance costs | 💰 Cost-effective operations |
| 📚 Steep learning curves | 🎯 Production-ready in minutes |
| 💤 Unknown resilience | 🛡️ Chaos-tested by the T-800 |
Timeline comparison: Days of prep versus minutes to prod.
🚀 Powerful Use Cases
- Voice-First Companions: Scarlett (voice agent) plus Daredevil (LiveKit) deliver low-latency, human-grade conversations.
- Stateful Research Agents: Sherlock (LangGraph) and Oracle (Postgres) keep long-running investigations grounded in truth.
- Adversarially Tested Automation: Ivan Drago (Gym) and the T-800 (Chaos Mesh) stress-test logic and infrastructure before production.
- Scalable AI Platforms: Use A.R.C. as the foundation for your own AI product suite across microservices and workers.
📚 Get Up and Running in Minutes
Dive into the documentation to scaffold your first platform and start composing agents immediately.
arcCLI walkthroughs in the introduction docs.- Service-by-service breakdown in
SERVICE.MD. - End-to-end platform guides in
docs/stack. - Source code and community support on GitHub.
Ready to build? Run arc new, follow the prompts, and launch with arc run. Your production-ready agent platform is one command away.
dgtalbug
DgtalBug