OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
Agent Frameworks

OM1

OM1 is an open-source, modular AI runtime and hardware abstraction layer for building multimodal agents across digital environments, phones, robots and simulators. It connects inputs such as web data, camera feeds and LIDAR to language or vision-language models and to outputs including speech, robot motion and autonomous navigation.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

OM1 — OpenViST project profile in Agent Frameworks.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use OM1 when developing an agent that needs a common runtime across multimodal inputs and digital or physical actions, especially when robot hardware connections or simulator-based prototyping are part of the project.

Who it is for

  • Developers building multimodal agents for robots or digital environments.
  • Robotics teams connecting agent pipelines to hardware through plugins and middleware.
  • Developers prototyping robot-agent behavior in Gazebo or Isaac Sim.

Documented uses

  • Combine camera, microphone, web or sensor input with model processing and agent responses.
  • Connect robot-specific hardware through ROS 2, Zenoh or CycloneDDS integrations.
  • Prototype conversational interaction and physical actions such as motion and autonomous navigation.
  • Monitor pipeline measures such as LLM and ASR latency with the included Prometheus and Grafana stack.
Before you choose

Requirements & limits

The repository is public and specifies an MIT license. Its quick start requires an OpenMind API key. Prebuilt releases are listed for Linux and macOS architectures; source builds require Go 1.25.0 or later and make.

  • The Go runtime is recommended for most new development; the available Python runtime is deprecated and will not be maintained.
  • The Go runtime covers the core agent pipeline, while some capabilities from the Python runtime are still under development.
  • OM1 supports ROS 2, Zenoh and CycloneDDS hardware connections, with Zenoh recommended for new development.
  • The conversation-agent quick start depends on audio system libraries and permissions for webcam use.
  • Grafana and Prometheus monitoring is an optional Docker-based component.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Set up and launch the conversation agent

Prepare OM1’s documented conversation-agent example for voice and camera interaction.

Before you start

  • A supported Linux or macOS system; listed prebuilt binary targets are linux-amd64, linux-arm64, darwin-arm64 and darwin-amd64.
  • PortAudio and FFmpeg system dependencies; Linux setup also calls for pkg-config.
  • An OpenMind API key.
  • A webcam and microphone for the documented voice-and-camera interaction.

Steps

  1. Install the documented system dependencies: on macOS, brew install portaudio ffmpeg; on Linux, install portaudio19-dev, ffmpeg and pkg-config.

  2. Download and extract a prebuilt release for a listed platform, or clone the repository and build from source with Go 1.25.0 or later and make.

  3. Set OM_API_KEY in the shell environment, or use a project-local .env file when building from source.

  4. Launch the conversation configuration using the platform-specific library path shown in the repository for a prebuilt binary, or run CONFIG=conversation make run for a source build.

Expected outcome

The conversation agent starts and can process webcam and microphone input, with responses spoken aloud; startup and pipeline activity are shown in the terminal.

Guide limitations

  • The documented prebuilt targets are limited to the listed Linux and macOS architectures.
  • Camera use requires granting the terminal or IDE camera permission in the operating system.
  • A source build requires Go 1.25.0 or later and make.
Documentation 1 ↗
Source-based practical guide

Enable OM1 latency monitoring

Start the included Prometheus and Grafana stack for viewing AI pipeline metrics.

Before you start

  • Docker installed.
  • The OM1 repository or its supplied runtime assets.

Steps

  1. Run docker-compose up -d grafana prometheus.

  2. Open http://localhost:3000 in a browser to view the provisioned OM1 Latency Monitoring dashboard.

Expected outcome

The dashboard provides real-time pipeline metrics, including LLM and ASR latency.

Guide limitations

  • The repository describes this monitoring setup as optional and requires Docker.
  • The documented default Grafana login is admin/admin.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
10 October 2026
Original website title
OpenMind/OM1 · Official homepage title · retained in its source language
How we review sourcesSuggest a correction