# Edge Impulse — AI Compare Ninja

> Collect sensor data, train and optimise edge ML models, and deploy them to embedded devices.

## Overview
- **Category:** Development & Code (Edge AI & embedded ML)
- **Status:** Active · checked - Active Developer/Enterprise structure. Professional officially replaced.
- **Paid from:** contact sales / Contract (Enterprise)
- **Free access:** Free plan (Three private projects, 10 experiments/project, 60 min compute/job.)
- **Website:** https://www.edgeimpulse.com/
- **Canonical URL:** https://www.aicompare.ninja/en/tool/edge-impulse

## What it does
- Collect, label, and train TinyML (Tiny Machine Learning) models in the cloud
- EON (Edge Optimized Network) compression reduces model size for microcontrollers
- Deployment packages for Arduino, Raspberry Pi, STM32, and other embedded target platforms
- DSP (Digital Signal Processing) and performance testing directly in the browser
- Free Developer plan with no time limit; Enterprise on request

## In brief
Edge Impulse is a development platform for TinyML and Edge AI. It combines data collection, labeling, DSP preprocessing, model training, EON size optimization, performance testing, and deployment packages for microcontrollers and edge hardware in a browser-driven workflow. The Developer plan is fully featured and free without time limits; Enterprise plans for more resources and commercial deployments are available on request.


## Limits
Developer publishes no total monthly compute pool, only a 60-minute maximum per job; ten experiments/project and three private projects add constraints. GPU minutes count three times CPU minutes when a contract includes a total allowance. Production and third-party distribution require the appropriate Enterprise Production Phase. The old $1,500/month and $400 Professional claims are obsolete.

## Pricing tiers
| Tier | Price | Period | What you get |
| --- | --- | --- | --- |
| Developer | free | Permanent | Three private and unlimited public projects; 10 experiments/project; Up to three collaborators/project; 60 minutes compute/job and 16 GB CPU RAM; Access to GPU training and GenAI assisted labelling; 30 minutes synthetic calibration data; Internal R&D/pre-production |
| Enterprise | contact sales | Contract | 10+ private projects and enterprise collaboration; Unlimited job duration and configurable memory; Full API access instead of Limited, RBAC, SSO, and white label; 120 minutes calibration data; Custom DSP/ML, support, and production licensing |

## Sources
All details regarding plans, limits, and features are based on official vendor documentation and the primary sources verified below.

- https://www.edgeimpulse.com/pricing
- https://www.edgeimpulse.com/blog/announcing-professional-plan/
- https://www2.edgeimpulse.com/faqs

---
*Source: AI Compare Ninja (2026) - https://www.aicompare.ninja/en/tool/edge-impulse*