Public Beta — May 2026

The Control Plane for
AI Coding Agents

Connect, route, monitor, and govern every AI coding agent your team uses — Claude Code, Codex, Hermes, OpenCode, and more — from a single dashboard.

app.agentforges.dev/dashboard
AGENTS CONNECTED
4
+2 this week
MONTHLY COST
$1,247
↓ 38% with routing
AGENT SESSIONS
2,847
this month
SAVINGS TO DATE
$4,210
since install
Connected Agents Last 30 days
CC
Claude Code
anthropic/claude-opus-4-5
$612
843 sessions
CX
OpenAI Codex
openai/gpt-5
$389
612 sessions
HM
Hermes Agent
deepseek/deepseek-v4-pro
$156
892 sessions
OC
OpenCode
anthropic/claude-sonnet-4-5
$90
500 sessions

Integrated with every major AI coding agent

Claude Code OpenAI Codex Hermes Cursor CLI OpenCode Copilot

Your Team Runs Multiple Agents.
Nobody's Managing Them.

Until now. AgentForges gives you a single pane of glass for every AI coding agent in your organization.

The Chaos

Each agent has its own auth, billing, and logs. Your Claude Code spend is invisible to finance. Your Codex sessions are untracked. You're flying blind with thousands in monthly AI costs.

The Control Plane

One CLI connects all your agents. One dashboard shows all costs, sessions, and effectiveness. One policy engine enforces governance everywhere. af connect and you're done.

The Outcome

40-60% cost reduction with intelligent routing. 100% visibility into agent activity. Zero surprises with Git-native approval workflows. Your agents work harder, cost less, and never go rogue.

Everything You Need to
Tame Your Agents

Five core features that turn agent chaos into engineering velocity.

One-Command Connect

af connect auto-discovers every AI coding agent on your machine — Claude Code, Hermes, Codex, OpenCode, Cursor, and more. No config files. No YAML. 30 seconds to onboard.

Cross-Agent Cost Dashboard

Real-time cost tracking across every agent, model, and developer. See exactly where your AI budget goes — with anomaly alerts when an agent goes into a spending loop.

ModelMatch™ Intelligent Routing

Our classifier analyzes each prompt and routes it to the optimal model — fast/cheap for simple tasks, powerful for complex ones. Save 40-60% without sacrificing quality.

Git-Native Approval Workflows

Policy-as-code that blocks agents from pushing to main without PR approval, modifying infrastructure without SRE sign-off, or making 50+ file changes without human review. Enforced at the agent level.

Agent Win Rate Analytics

Git-integrated metrics that track which agents produce accepted code vs. rework. "Agent Win Rate" = accepted PRs / total agent PRs. Data-driven decisions on which agents to standardize on.

Simple, Transparent
Pricing

Start free. Upgrade when your team needs visibility and governance.

Team

For engineering teams that need visibility and cost control.

$29 /dev/mo
  • Unlimited AI coding agents
  • Cross-agent cost dashboard
  • ModelMatch™ intelligent routing
  • 90-day analytics history
  • Team analytics & benchmarks
Start Free Trial
POPULAR

Business

For organizations that need governance and compliance.

$49 /dev/mo
  • Everything in Team
  • Git-native approval workflows
  • SSO (SAML/OIDC) & RBAC
  • Audit logs & compliance exports
  • Custom routing policies
Start Free Trial

All plans include the open-source CLI (AGPL). Enterprise plan available for on-prem deployment, dedicated tenancy, and custom SLAs. Contact us.

What Early Users
Are Saying

"We were running Claude Code for architecture, Codex for CI/CD, and Cursor for daily dev — and had zero idea what we were spending. AgentForges's dashboard showed us we were burning $2,400/month unnecessarily. The ModelMatch router cut our costs by 43% in the first week."

RS
Rafael Silva
VP Engineering, Series B SaaS

"Our security team was terrified of AI agents touching production. AgentForges's approval workflows gave us the guardrails we needed — agents can't push to main without PR approval, period. We got the velocity without the risk."

AK
Amanda Kowalski
CISO, Fintech Unicorn

"I was skeptical about adding another tool to our stack. But `af connect` took literally 30 seconds, and suddenly I had a dashboard showing all our agent costs. My CFO loves the monthly report. My team loves that they don't have to think about which model to use anymore."

MP
Marcus Pereira
Head of Platform, Scale-Up

Ready to Tame
Your AI Agents?

One command. All your agents. Total control.

$ curl -sSL https://agentforges.dev/install | sh

Then: af connect → dashboard live in 30 seconds. No credit card required.