Back to directory

TitanX

An enterprise desktop application for orchestrating teams of AI agents, adding audit logging, tool-access policies and observability on top of an existing open-source multi-agent chat interface.

TitanX

TitanX is built on the open-source AionUI platform and adds an identity and access management layer, workflow automation inspired by n8n, agent-memory features inspired by LangChain and DeepAgents, and LangSmith-compatible tracing, turning what its own documentation describes as a multi-agent chat interface into a wider agent-management control plane. Agents from providers including Claude, Gemini and OpenCode work within it, alongside more than 20 LLM providers overall.

Described capabilities include an HMAC-signed, immutable audit log covering every action, four built-in security profiles called agent blueprints for sandboxed, developer, researcher and CI use, an n8n-inspired visual workflow builder with triggers, conditions and approvals, and a command center dashboard tracking key metrics, token usage and cost. Its architecture routes a chat, sprint board and other renderer surfaces through an IPC bridge with a channel whitelist to a backend team-orchestration and agent-OS layer, storing state in SQLite with a versioned set of migrations.

It is an Electron desktop application written in TypeScript and React, and is released under the Apache-2.0 license.

More in Agent frameworks

See all 902
  • Recurse

    recurse.run

    A serverless harness that lets a coding agent design, test, and deploy its own specialist sub-agents as tools, MCP servers, or bots, defined by a schema-validated manifest.

  • ProofForge

    shulgin.is-a.dev

    An AI-agent pipeline that decomposes a mathematics problem, proves each piece, and formalizes the result in Lean 4 and Mathlib until the Lean kernel accepts it, with six of its proofs already merged into Google DeepMind's formal-conjectures repository.

  • local-coder

    github.com

    Builds a local coding-agent team on top of Ollama and OpenCode by detecting the machine's hardware, recommending models for seven specialist roles, and testing whether each assigned model can actually use tools, edit a repository and pass verification.