Skip to main content

Saykai

Saykai provides a CI-based safety gate for teams deploying ROS 2 and Nav2 robots, running a deterministic three-room audit on every pull request. The audit checks secrets, behavioral simulations, and Nav2 policy against an approved Safety Spec, producing an Ed25519-signed Safety Pack as an immutable record. The system runs entirely within the customer's own CI pipeline, with GitHub Actions integration and shadow mode for initial rollout.

Traction –/10
HQ unknown
0+ followers
  • Robotics
  • Developer Tools
  • Regulatory & Compliance Technology
  • Industrial Automation
  • Software Only
Updated 4 days ago

Funding

Funding rounds are not available yet.

Founders

Founder details are not available yet.

Product

Problem

Teams deploying ROS 2 and Nav2-based robots lack a systematic way to verify that every code change meets safety requirements before it reaches production hardware. Manual review processes are inconsistent, and there is no immutable record of what was checked or approved, making it difficult to demonstrate safety compliance to regulators or internal stakeholders.

Solution

Saykai turns CI into the first safety line for robot deployments by running a deterministic three-room audit on every pull request. The audit includes a secret scan, behavioral simulation, and Nav2 policy check, with results compared against an approved Safety Spec defined in a saykai-policy.yml file. Each run produces a signed JSON Safety Pack that records every finding and the exact Safety Spec version used, providing a clear pass or block decision backed by Ed25519 signatures. The system runs entirely inside the customer's own CI pipeline, ensuring no data leaves their infrastructure, and integrates as a required check on GitHub Actions.

Target Audience

Primary customers are engineering teams deploying Physical AI, AMR fleets, and warehouse robotics on ROS 2 and Nav2, particularly those needing verifiable safety compliance in their development pipelines.

Features

  • Three-room audit pipeline: secret scan, behavioral simulation, and Nav2 policy check on every pull request
  • Policy enforcement via saykai-policy.yml with configurable hard limits for max speed, acceleration, and safety buffer
  • Ed25519-signed Safety Pack JSON output serving as an immutable audit trail for each release decision
  • Runs entirely within the customer's own CI environment with zero data leaving the pipeline
  • GitHub Action integration as a required check, with support for other CI systems during pilot
  • Versioned Safety Spec tracking that records the exact policy version used for each audit
  • Shadow mode operation that reports potential blocks without failing builds, enabling safe initial rollout
  • Optional recorded approval workflow for changes to safety limits, requiring a different approver than the last editor
This profile is AI-generated and may contain inaccuracies.