Features

What it takes to run
an agent for real.

ControlForge isn't a model wrapped in a loop. It's a deterministic execution runtime your agent calls to actually do the work — stateful, auditable, and batteries-included. Every number below is measured against the live runtime.

2,000+
built-in capabilities
2,247 across 95 categories
95
capability categories
files, data, code, comms…
1
static binary
pure Go, no CGO
MCP
native tool surface
plugs into your agent

The loop

The non-deterministic part never executes anything directly. Five stages turn a probabilistic proposal into a bounded, recorded action.

01

Propose

The agent introspects a typed tool surface — every capability it can call — and proposes the steps against what actually exists. No guessing, no hallucinated APIs.

  • Live capability introspection
  • 2,247 capabilities across 95 categories
  • MCP-native tool surface
02

Verify

Proposed logic is validated against the runtime before it runs. Malformed or unsupported steps are rejected at the door — they never touch a real system.

  • Validation before execution
  • Rejected logic never runs
  • Errors returned to the agent, not the world
03

Bound

Scheduling, role-based permissions, and explicit limits constrain what any step is permitted to do — and when.

  • Deterministic scheduling & priorities
  • Role-based access control
  • Explicit, declared side effects
04

Execute

A deterministic engine runs the steps — the same way every time. State persists across steps and restarts. Hot paths run on a compiled bytecode VM.

  • Deterministic execution
  • Durable state across runs
  • Bytecode VM fast path
05

Audit

Every value and action lands in the historian and event stream, time-stamped. A deterministic state hash and full diagnostics make any run replayable and provable.

  • Time-stamped historian
  • Event stream + alarms
  • Deterministic state hash

Capabilities

The substance under the story — what your agent can call, and what keeps it honest.

Agent interface

  • MCP-native tool surface
  • Live capability introspection
  • Validate-before-execute contract
  • Subscribe to ground truth via the event stream
  • Replay any run against the historian

Deterministic execution core

  • Deterministic execution engine
  • Durable state across steps & restarts
  • Bytecode VM compute fast path
  • Reproducible: same in, same out
  • A simple, inspectable step language

2,000+ built-in capabilities

  • Files, HTTP, WebSockets, SMTP
  • JSON, CSV, INI, regex, compression
  • Databases & key/value stores
  • Crypto, checksums, encoding
  • Math, stats, strings, datetime
  • Shell / process execution

State & audit

  • Built-in historian (time-stamped)
  • Event stream + alarms
  • Diagnostics: timing, fault tracking
  • Deterministic state hash
  • Live value streaming (SSE)

Scale & resilience

  • Multi-instance clustering
  • Failover handling
  • Store-and-forward buffering
  • Resilience primitives built in

Deploy anywhere

  • Single static binary, no CGO
  • Linux (amd64 / arm64)
  • macOS · Windows
  • Raspberry Pi & ARM boards
  • Runs on a laptop or a $15 SBC

It reaches the real world.

Most agent runtimes stop at the edge of the screen. ControlForge doesn't have to — when a task needs to touch a device, a sensor, or a machine, it already speaks the protocols. Not the pitch; just a door the others don't have.

ModbusEtherNet/IPOPC UAMQTTSparkplug BDNP3BACnetSNMPCAN / J1939 / OBD-IIGPIOArduino · RP2040 · TeensyDMX / ArtNet / sACNMIDI / OSCKNXGPS / NMEABarcode / ZPL

Give your agent something it can run on.

ControlForge is in private access. Tell us what you're building.

hello@controlforge.tools