Arca platform

The platform for AI agents with apps

Build intelligent solutions with the power of agents interacting with your custom apps. Arca provides the platform, so you can focus on your business.

Request a demo Discuss an Arca initiative

Arca platform — AI agent workspace with chat, research tools, and document analysis

What you get with Arca

Agent shell for time to market

Arca gives your LangGraph agents a complete product shell: UI, API, auth, observability, and evals without writing platform code. You register your agent as a plugin and ship.

Plug your agent into Arca

Drop any LangGraph compatible agent into Arca as a plugin and get the full stack: agent app UI, REST and WebSocket API, JWT auth, RBAC, session management, usage metering, and chat history.

Ship agent apps

Create collaborative workspaces. Ship an application UI alongside your agents so users stay in control and work with your agents focusing on outcomes.

Close the loop with evals

A built-in evaluation framework covering business outcomes, technical performance, accuracy, and risk. Close the delivery loop and ship agents with measurable, repeatable quality metrics.

Arca platform capabilities

Engineered to support your business

Chat UI

A production-grade chat interface for your users, out of the box. White-label, ready to deploy under your brand.

Agent app

Extend the workspace with your own React components. Define controls, visualizations, or any interaction your agent needs, rendered alongside the chat as part of your plugin.

REST + WebSocket API

Register your agent once and it's callable via standard REST and WebSocket endpoints. Ready to integrate with existing systems or expose to users.

Auth & session management

JWT authentication, RBAC, and refresh token handling come with the platform. Multi-user, multi-agent, session isolation.

Chat history

Conversation persistence per user, per agent. Users pick up where they left off.

Observability

Agent runs are tracked, inspectable, and comparable. See what happened, when, and why.

User memory (on the roadmap)

Arca maintains memory across sessions and passes it to your plugin. Your agent knows who it's talking to.

Usage metering (on the roadmap)

Define limits per user, Arca enforces them. Track who uses what and how much.

Evals

Define quality benchmarks and run them locally or in CI. Business outcomes, technical performance, accuracy, and risk are measurable on each release.

Get started with Arca in 3 steps

From plugin to production

Arca is self-hosted. Configure your agent as a plugin and add it to your Arca instance.

Bring your agent

Start with an existing LangGraph agent or build a new one. Optionally add your Agent App UI as part of the plugin.

Register via arca.json. A minimal agent runs as-is; incremental adaptation unlocks the full range of Arca platform services.

Configure and start Arca

Arca is self-hosted and runs entirely on your infrastructure. Pair it with on-premises models and agents, and your entire AI stack operates locally — with no data leaving your environment.

Arca is designed for teams building their next agent app

Support your business with Arca

Ship fast

Teams that need to ship agents and agent apps to production fast.

Prototype and innovate

Teams exploring new agent app ideas and prototyping at speed.

Data confidentiality

Organizations that keep their data confidential and on their own infrastructure.

Ready to ship your agent app?

Arca is currently in early access. Request a demo to get started.

Request a demo