Open Source · MIT Licensed

Write agents in AXON.
Govern them with Mesh.

Two open-source projects that work together as a unified agent development pipeline. Define AI agents in a typed DSL, compile to code, then run them through a 9-gate governance workflow before production.

⚡ Try Live Playground Quick Start How It Works AXON Repo Mesh Repo

The Pipeline

From .ax source to governance decision in one click.

📝
Write .ax
Define agents, tools, RAG, flows in AXON DSL
⚙️
Compile
Parse, validate, type-check, codegen to Python/TS/Go/Rust
📋
Govern
Submit to Mesh 9-gate governance workflow
✅
Deploy
Pilot or production-ready certification
# 1. Write an agent in AXON agent SupportBot { model: @mock/gpt tools: [Search, Reply] fn run(q: Str) -> Str { act Search(query: q) } } # 2. Compile and submit to governance axon govern support_bot.ax --mesh-url http://localhost:8000 # 3. Mesh runs 9 gates → decision: pilot_candidate
AXON

AXON DSL

A typed programming language where agents, tools, memory, RAG, and orchestration are first-class language constructs — not framework boilerplate.

  • Parse, validate, type-check .ax source files
  • Codegen to Python, TypeScript, Go, Rust, MCP servers
  • Web playground with live parse/validate/codegen
  • WASM parser for client-side processing
  • CLI with 20+ commands (parse, validate, govern, ci-template)
  • Distributed mesh runtime with Redis/NATS backends
  • Dashboard, debugger, profiler, trace replay
→ github.com/annapurnaagenticsolutions/axon
AgentOps Mesh

AgentOps Mesh

An open-source AgentOps control plane for governing, evaluating, operating, and auditing enterprise AI agents before production.

  • 9-gate deterministic governance workflow
  • Risk classification and autonomy scoring
  • Policy-as-code guardrails (allowed/blocked actions)
  • RBAC security with roles, tenants, capabilities
  • Observability trace ledger and audit event bus
  • Connector sandbox for tool execution
  • Provider gateway with model routing controls
  • 30+ REST API endpoints, static control-plane consoles
→ github.com/annapurnaagenticsolutions/open-enterprise-agentops-mesh

9-Gate Governance Workflow

Every AXON agent passes through these gates before it can reach production.

G1
Use Case Intake
Ownership, scope, and business context defined
G2
Suitability
Business value vs. task suitability for AI
G3
Risk Classification
Autonomy level, risk factors, required controls
G4
Data Readiness
Data quality, retrieval, lineage, freshness
G5
Evaluation
Reliability, safety, cost benchmarks
G6
Policy & Guardrails
Policy-as-code rules for allowed/blocked actions
G7
Security & RBAC
Roles, tenants, capabilities, access boundaries
G8
Audit & Traceability
Decision history, trace ledger, evidence vault
G9
Production Readiness
Pilot-ready → production-ready certification

Quick Start

Run both services together with one command.

# Clone both repos into the same directory git clone https://github.com/annapurnaagenticsolutions/axon.git git clone https://github.com/annapurnaagenticsolutions/open-enterprise-agentops-mesh.git # Run both services together docker compose -f docker-compose.unified.yml up --build # Open the playground → write an agent → click Govern # http://localhost:8080 — AXON Playground # http://localhost:8000 — AgentOps Mesh API # http://localhost:8081 — Mesh static consoles
⚡ Try Live Playground AXON Repo Mesh Repo Mesh Control Plane Governance Console