nff is an autonomous operator platform for embedded firmware fleets that lets engineers and AI agents build, deploy, and monitor device software through natural-language prompts. The platform automates the entire workflow, from writing and compiling firmware to flashing devices, debugging failures, and shipping over-the-air updates, giving users a single agent that runs the whole loop. It integrates with coding agents like Open Code, supporting use cases such as implementing BLE features on ESP32s or tuning MPC controllers on drones via MQTT.
Funding
Funding not disclosed
Founders
Product
Problem
Embedded software development is slowed by manual, repetitive steps such as bench reflashing, staged over-the-air (OTA) updates, and analyzing crash dumps, which fragment the engineering workflow. Engineers must constantly switch between coding, building, flashing, and debugging tools, making it difficult to iterate quickly on firmware for hardware fleets.
Solution
nff provides a fleet platform that pairs a coding agent with a device build, bench, and ship pipeline, compressing the entire firmware lifecycle into a single natural-language command. The agent plans and writes embedded firmware, compiles it, flashes it to connected hardware, and iteratively debugs errors until the code works. Once approved, the same agent opens a pull request and ships the update OTA to the fleet, then monitors devices in operation for anomalies. This autonomous loop removes manual handoffs between tools and lets engineers delegate the build-test-fix cycle while retaining final merge control.
Target Audience
Primary users are embedded software engineers, robotics developers, and hardware product teams who manage fleets of connected devices and want to integrate AI agents into their firmware build, test, and deployment workflows.
Features
- Natural-language workflow orchestration where a user prompt triggers the agent to implement, flash, debug, and ship firmware without manual intervention
- End-to-end pipeline covering device build, bench flash, staging, and post-release operation monitoring
- Integration with coding agents like Open Code to plan and implement features, open PRs, and ship OTA updates
- Automated error reproduction and root-cause analysis for issues such as startup spikes or motor failures, with the agent patching the code until it works
- Support for over-the-air (OTA) delivery and runtime retuning of device parameters via MQTT for connected fleets
- Local CLI option via `pip install nff` and a hosted platform for team-scale deployment