Blue Iris and Frigate are the two leading self-hosted NVR software packages in 2026. Blue Iris runs on Windows. Frigate runs in Docker on Linux, Synology, or a Home Assistant OS box. Both replace a hardware NVR with a PC + software setup that often costs less and offers more flexibility.
Quick Verdict
- Blue Iris: best for Windows users, broad camera support, mature stability, $69.95 one-time license.
- Frigate: best for Home Assistant integration, real-time AI object detection on a Coral TPU, free and open source.
- Pick Blue Iris if you have an old Windows PC and want a polished, stable NVR with the widest camera compatibility.
- Pick Frigate if you already run Home Assistant or want native MQTT events for home automation.
Side-by-Side
| Spec | Blue Iris 5 | Frigate 0.14 |
|---|---|---|
| Platform | Windows 10/11 | Docker (Linux, Synology, HA OS, Unraid) |
| License | $69.95 one-time | Free, MIT license |
| Camera support | ONVIF + most proprietary protocols | RTSP + ONVIF |
| AI detection | DeepStack or CodeProject.AI (optional) | TensorFlow Lite (Coral TPU recommended) |
| Home Assistant integration | Via HACS custom component | Native MQTT + HA integration |
| Hardware needs | i5+ Windows PC, GPU for AI | Raspberry Pi 4 + Coral TPU works |
| Setup difficulty | Easy (Windows installer) | Medium (Docker, YAML config) |
When Blue Iris Wins
You have an old Windows PC. Blue Iris uses the OS you already know, the file system you already manage, and the camera-detection wizard handles 90 percent of brands automatically.
You need maximum camera compatibility. Blue Iris ships with templates for over 200 camera brands. Frigate handles fewer cameras directly (most are added as generic RTSP streams).
You want polished native UI. The Blue Iris Windows app has 20+ years of UI polish. Frigate’s web UI is functional but minimal.
One-time pay vs recurring infrastructure. $69.95 and you own it. No Docker maintenance, no YAML updates, no breaking changes.
When Frigate Wins
You run Home Assistant. Frigate is the de facto standard for HA users. Object detection events fire MQTT messages that trigger HA automations directly. Blue Iris needs an HACS integration that often lags behind HA core updates.
You want real-time on-device AI. Frigate with a Coral USB TPU ($60) processes object detection in under 20 milliseconds per frame. Blue Iris with CodeProject.AI runs slower (50-200ms) unless you have a strong NVIDIA GPU.
You prefer Linux or containers. Frigate runs on a $80 Raspberry Pi 4 with a Coral TPU. Blue Iris needs a Windows PC with at least 8 GB RAM.
You want open source. Frigate’s code is on GitHub. If a feature breaks, you can read the source. Blue Iris is closed source.
Hardware Recommendations
Blue Iris on Windows
8 GB RAM minimum, i5-9400 or newer CPU. Intel Quick Sync handles H.265 decoding for 8 to 16 cameras without a discrete GPU. Add a $200 NVIDIA RTX 3050 if you want CodeProject.AI to run at full speed.
Storage: dedicated 4 to 8 TB Western Digital Purple drive separate from the OS drive.
Frigate on Synology NAS
Synology DS923+ (4-bay, $580 without drives) runs Frigate in Container Manager smoothly for 4 to 8 cameras. Add a Coral USB TPU for AI detection. Best path for users who already own a Synology NAS.
Frigate on Home Assistant Yellow
HA Yellow with a Coral PoE add-on board ($85) runs Frigate alongside Home Assistant on one device. Excellent for HA-first households; limited to about 6 cameras at 1080p.
AI Detection Compared
Both tools detect persons, vehicles, animals, and packages. Detection accuracy is similar (95 to 98 percent in good lighting). The differences:
- Frigate with Coral TPU: 10 to 20 ms per frame, all cameras processed in parallel.
- Blue Iris with CodeProject.AI on CPU: 100 to 300 ms per frame.
- Blue Iris with NVIDIA GPU: 20 to 50 ms per frame.
Frigate’s TPU advantage matters when you have 6 or more cameras streaming simultaneously and want immediate notifications. For 4 or fewer cameras the difference is invisible to users.
Home Assistant Integration
Frigate sends MQTT events that HA automations consume natively. A motion event with a person object on the porch camera can flash a Hue light, announce on a Nest speaker, and enable the door for a known face, all from one HA automation.
Blue Iris uses an HACS custom component that exposes camera states to HA. It works but the automation surface is narrower (cannot trigger on specific detected objects without extra glue scripts).
FAQ
Can I run both?
Yes. Some users run Blue Iris for full recording and Frigate solely for AI event detection. The cameras feed both via RTSP. This setup uses double the bandwidth but the strongest of each app.
Will Blue Iris run on a Mac?
No. Blue Iris is Windows only. Mac users typically run Blue Iris in Parallels or pick Frigate.
Does Frigate need a Coral TPU?
Not required but recommended. Without a TPU, Frigate uses CPU for detection and slows down past 4 cameras.
Which is cheaper over 5 years?
Frigate: $0 license, $60 Coral TPU, $80 Raspberry Pi = $140. Blue Iris: $69.95 license + existing Windows PC = $70 + electricity. Hardware cost is similar; Blue Iris wins on simplicity.
Related Guides
- Blue Iris vs Synology Surveillance Station
- What is an NVR?
- Best AI NVR Hardware
- Motion Detection Guide
- Storage Calculator
Related Terms and Concepts
Key topics linked to blue iris vs frigate: integrate, google coral, alert, Central processing unit, Docker (software), Computer hardware, Microsoft Windows.
Adjacent technical terms and brand context: Google, Personal computer, Artificial intelligence, Graphics processing unit, Computer, Doorbell, Graphical user interface, Server (computing).
Common Questions and Related Topics
Configuration File
Configuration File is a common topic linked to blue iris vs frigate. Buyers comparing options often look at unifi, using blue iris, home security when evaluating fit.
Real-Time Streaming Protocol
On the question of Real-Time Streaming Protocol: most setups for blue iris vs frigate touch this area. Specs to check include home security, best nvr, configure and how they apply to your install.
Ecosystem
Ecosystem shows up in nearly every blue iris vs frigate comparison. Pay attention to configure, continuous recording, false positives before deciding.
Blueiris
Blueiris is a common topic linked to blue iris vs frigate. Buyers comparing options often look at false positives, playback, frigate and blue iris when evaluating fit.
Reolink
On the question of reolink: most setups for blue iris vs frigate touch this area. Specs to check include frigate and blue iris, frigate vs, like blue iris and how they apply to your install.
Use Frigate
Use Frigate shows up in nearly every blue iris vs frigate comparison. Pay attention to like blue iris, frigate seems like, integration with home assistant before deciding.
More Topics to Consider
Additional blue iris vs frigate concepts worth reviewing: integration with home assistant, smart home, setup advice, nvr systems, frigate can handle, shinobi.
Specifications and Buying Notes for Blue Iris Vs Frigate
Hardware Recommendations Compared
Both apps run on completely different hardware profiles. Sizing matters more than software preference.
Blue Iris baseline
Minimum: Windows 10/11, Intel i5-8400 or AMD Ryzen 5 3600, 8 GB RAM, Intel Quick Sync iGPU for H.265 decoding, 4 TB Western Digital Purple drive separate from the OS drive. This handles 8 to 12 cameras at 1080p comfortably.
For AI: add a $200 NVIDIA RTX 3050 6 GB. With CodeProject.AI loaded, the GPU handles person and vehicle detection in 20-50 ms per frame across all cameras simultaneously.
Frigate baseline
Minimum: any Linux box with Docker (Raspberry Pi 4 8 GB works for 4 cameras, an Intel NUC 11 for 8+), 4 GB RAM minimum, USB Coral TPU ($60) for AI inference. Storage on a separate USB-attached SSD or NAS.
Best mid-tier setup: Synology DS923+ NAS ($580 without drives) running Frigate in Container Manager + Coral USB TPU. Handles 8 cameras at 1080p with sub-30ms AI inference per frame.
AI Inference Speed Side-by-Side
| Setup | Time per frame | Cameras supported |
|---|---|---|
| Frigate + Coral USB TPU | 10-15 ms | Up to 16 at 1080p |
| Frigate + Coral M.2 dual TPU | 5-8 ms | Up to 32 cameras |
| Blue Iris + CPU only (i5) | 100-300 ms | Best for 4-6 cameras |
| Blue Iris + NVIDIA RTX 3050 | 20-50 ms | 8-12 cameras |
| Blue Iris + NVIDIA RTX 4060 | 10-25 ms | 12-20 cameras |
Frigate with a Coral TPU is dramatically faster than Blue Iris CPU-only. Once Blue Iris has a discrete NVIDIA GPU, the gap closes.
Home Assistant Integration Depth
This is where Frigate has its clearest win for HA users:
Frigate + HA
Native integration via the Frigate HACS component. Each AI-detected object fires an MQTT event that HA picks up within 100 ms. Automation example: ‘When Frigate detects person in zone front_porch AND sun is below horizon, turn on Hue porch light and play TTS welcome on Nest Mini.’ Trigger latency under 1 second.
Blue Iris + HA
Integration via the Blue Iris HACS component. Camera states and motion events appear as binary sensors that HA polls every 5-15 seconds. Same automation works but with 5-15 second trigger latency, not instant.
For latency-sensitive automations (porch lights triggered by motion), Frigate is significantly better. For pure security recording, the difference is invisible.
License and Cost Over 5 Years
| Item | Blue Iris | Frigate |
|---|---|---|
| Software license | $69.95 one-time | Free (MIT) |
| Required hardware | Existing Windows PC | Raspberry Pi 4 8 GB ($80) or Synology NAS |
| AI accelerator (optional) | RTX 3050 ($200) for fast AI | Coral USB TPU ($60) recommended |
| OS license | Windows already paid | Free (Linux) |
| Total 5-year hardware cost | ~$269 if buying GPU | ~$140 for Pi + Coral |
| Support model | Active developer (Ken Pletzer) | Open-source community + GitHub |
| Update cadence | Monthly Windows updates | Quarterly Docker image releases |
Frigate wins on raw cost for AI-heavy setups. Blue Iris wins on per-camera simplicity if you already have a Windows PC sitting idle.
Migration Path Between Them
Both tools can coexist on the same network. Two common patterns:
Frigate for AI events, Blue Iris for recording
Cameras stream the same RTSP feed to both apps simultaneously (modern cameras support 3-5 concurrent connections). Frigate handles object detection and fires Home Assistant automations. Blue Iris handles 24/7 recording and timeline scrubbing. Best of both worlds; doubles bandwidth on the LAN.
Migration from Blue Iris to Frigate
Export camera RTSP URLs from Blue Iris (Settings > Cameras > Export). Import into Frigate’s YAML config under cameras: section. Frigate’s first-pass auto-detect runs within 24 hours of bringing the cameras online. Disable Blue Iris recording once you confirm Frigate is capturing what you need.
Migration from Frigate to Blue Iris
Less common direction but doable. Export Frigate camera YAML, manually add each camera in Blue Iris’s Add Camera wizard. Blue Iris will auto-detect codec, resolution, and PTZ support via ONVIF.
Real-World Verdict by User Type
- Windows user, 4 to 8 cameras, no Home Assistant: Blue Iris. Simpler setup, polished Windows UI.
- Home Assistant power user, any camera count: Frigate. The MQTT integration is unmatched.
- Synology NAS owner: Frigate via Container Manager. Reuses existing hardware.
- Raspberry Pi tinkerer: Frigate + Coral TPU. Cheapest possible AI NVR setup.
- 10+ cameras, mostly recording-focused: Blue Iris with NVIDIA GPU. More mature playback and clip review UI.