Plans that scale from prototype to fleet.
Start free. Upgrade when your robot is more than a side project. Pay annually for two months free.
Free
For curious developers and students.
- Basic AI chat
- Small context window (32k)
- Limited ROS workspace indexing
- Community support (Discord)
- Basic code completion
- 5 simulation runs / month
- Single workspace
Pro
For roboticists shipping real machines.
- Advanced AI robotics assistant (Sonnet 4.6)
- Full ROS2 workspace intelligence
- TF tree analysis & QoS auditing
- Launch file symbolic execution
- Sim-in-the-loop debugging
- Bag file analyzer (MCAP + rosbag2)
- Embedded support (STM32 / ESP32 / Zephyr)
- Priority inference, larger context (200k)
- Unlimited workspaces
- Email support
Max
For teams operating fleets.
- Autonomous robotics agents (Opus 4.7)
- Multi-agent debugging loops
- Cloud Gazebo / Ignition GPU pool
- Team collaboration & shared RKG
- Fleet debugging & remote agents
- Enterprise deployment tools
- Priority inference + dedicated compute
- Custom robotics fine-tunes
- Advanced observability & audit
- API access
- SLA & dedicated support
Need 50+ seats, on-prem, or air-gapped? Talk to our enterprise team →
Compare plans
Every capability, side by side.
| Feature | Free | Pro | Max |
|---|---|---|---|
| AI | |||
| Frontier model | Haiku 4.5 | Sonnet 4.6 | Opus 4.7 |
| Context window | 32k | 200k | 1M (cached) |
| Autonomous agents | |||
| Multi-agent debugging | |||
| Custom fine-tunes | |||
| Robotics | |||
| ROS2 workspace indexing | limited | ||
| Launch IR & TF analysis | |||
| Bag analyzer | 1 bag | unlimited | unlimited + cloud |
| Embedded (STM32/ESP32/Zephyr) | |||
| Fleet / remote robot agent | 1 robot | unlimited | |
| Simulation | |||
| Local Gazebo / Ignition | 5/mo | unlimited | unlimited |
| Cloud GPU sim | 20 hrs/mo | 200 hrs/mo | |
| Scenario library | yes + custom | ||
| Team & enterprise | |||
| Workspaces | 1 | unlimited | unlimited |
| Shared RKG | |||
| SSO / SAML | |||
| Audit logs | |||
| Self-hosted option | |||
| Support | Community | Dedicated + SLA | |
Frequently asked
Do I need a GPU?
No. RoboAgent runs on a regular Linux laptop. GPU is only useful for cloud Isaac Sim (Max tier) or self-hosted local LLMs.
Does it work with my existing ROS2 workspace?
Yes. Point RoboAgent at your ros2_ws and it will index colcon, launch files, URDF, and bags incrementally. No code changes required.
Can I use my own LLM keys?
On Pro and Max, yes — bring-your-own-key for Anthropic, AWS Bedrock, or Vertex. Useful for compliance and bursty enterprise inference.
How is data handled?
Source code and bags stay local by default. Telemetry to improve our models is opt-in only. Enterprise customers can run fully air-gapped.
Can I cancel anytime?
Yes. Monthly plans cancel instantly; yearly plans are refundable for 30 days, no questions asked.
Is there an academic discount?
Pro is free for verified .edu accounts. Reach out from your institutional email and we'll set you up.
30-day refund policy
If RoboAgent doesn't save you a meaningful amount of debugging time in the first 30 days, email billing@roboagent.ai for a full refund. No invoices, no hoops. Yearly plans are pro-rated to the day.