Start free

AnyForge Control

Govern every AI call.Change nothing else.

AnyForge Control is the governance layer underneath the AnyForge platform. For teams using their own agents (Claude Code, Cursor, Cline, Continue, or a raw SDK client), it ships as a governed LLM proxy that installs in five minutes. Zero workflow change. Immediate visibility.

Start free See the install steps

Using AnyForge Code Studio or AnyForge Crew? Control is already native there, with nothing separate to install. This page is for the “bring your own agent” path.

4-step install · 5 minutes

Four steps to your first governed call.

  1. Sign up at anyforge.ai. Free, no card. You get a 10M-token trial on signup.
  2. Add your provider API keys at /settings/secrets: Anthropic, OpenAI, Google or OpenRouter. Keys live in Google Secret Manager, never in our database.
  3. Run the CLI in your repo. It detects your agent, mints a scoped AnyForge API key and writes the proxy config.
  4. Open your dashboard at crew.anyforge.ai: cost analytics, routing decisions, audit chain. Your next LLM call shows up within seconds.
# In your repo
$ npx @anyforge/control init
  • Claude Code
  • Cursor
  • Cline
  • Continue
  • Raw SDK client

After the trial, you pay one flat, provider-agnostic platform fee. The rate is the same whether you use cloud LLMs or self-host. No subscription, no seats, no daily caps. Routing and caching are projected to save more on wasted model spend than the fee costs, and your Usage page shows what caching saved you each month.

Capabilities

Visibility, routing and governance. Opt-in per rule.

  • Unified proxy

    One endpoint in front of these providers. SDK-compatible, with no prompt rewrites.

    • OpenAI
    • Anthropic
    • Gemini
    • OpenRouter

  • Self-hosted Ollama

    Run models on your own Ollama daemon, with the same governance on every call, including air-gapped, data-residency and regulated deployments.

    • Local
    • On-prem
    • Ollama Cloud

  • Cost analytics

    Per-call, per-agent and per-project spend, with dimensional rollups and alerts at the thresholds you set.

  • Intent-based routing

    Rules, regex or embedding-based routing. Each call goes to the cheapest model that meets the quality bar.

  • Portable memory

    Shared context that survives provider switches. Same conversation history, different model, no rebuild.

  • Hash-chained audit

    A signed, append-only log of every prompt, response and tool call. Regulator-ready.

  • Access controls

    Each call is checked against your IP allowlist and the key holder’s role.

    • IP allowlists
    • Role-checked keys

  • Claude Max routing (OAuth)

    Route agent traffic through your subscription with an sk-ant-oat-… token, not metered API rates.

Already on Claude Max?

Route through your subscription, not a metered API account.

Per-token API rates add up fast on agent workloads. AnyForge accepts two kinds of Anthropic credential in Settings → Providers → Anthropic.

Docs › Claude OAuth tokens (beta) for setup.

CredentialHow it is billed
sk-ant-api-…Standard API key, per token at Anthropic's API rates.
sk-ant-oat-…OAuth token from claude setup-token, billed against your Claude Max or Pro subscription.
  • Flat quota. The Max quota absorbs the same traffic at no marginal cost, within Anthropic's rate limits.
  • Fallback key. An optional slot keeps the harness running if Max saturates. It falls through to a metered key, with a notice on /usage.

What you learn on Control

Three questions the data answers in the first week.

  • What did this feature cost in tokens?

    Per-call spend, grouped by agent, project or user.

  • Which model pays for itself on this kind of task?

    Latency, success rate and cost per outcome, model by model. Pick the cheapest model that meets the quality bar, with evidence.

  • Which workflows are ready for multi-agent delivery?

    Repeated tasks with real stages, dependencies and a human approver show up as Crew candidates on the dashboard.

When you are ready

Graduate to AnyForge Crew. Keep Control running underneath.

Crew is governed multi-agent delivery, and you turn it on per project. Control keeps running underneath, so spend and audit stay unified across both planes.

You do not have to move to Crew. Many teams run Control indefinitely, and that is a valid end state.

Start a crew Back to overview
  • One Engineer drives the build. Architect, Security and QA specialists advise on demand, and a parallel specialist review checks the PR diff.
  • One mandatory human approval gate at PR review. The one opt-in exception: Release Health auto-merges small safe-path hotfix PRs without operator review when you enable it.
  • A cryptographic audit trail on every call.

Install Control in five minutes.

Free to start, on the provider keys you already have. No seats and no workflow change.

Start free See pricing