10 Jan 18:23

blakeblackshear

93016c6

0.17.0 Beta 2 Pre-release

Pre-release

Beta docs: https://deploy-preview-19787--frigate-docs.netlify.app/

Images

Changes since Beta 1

Restrict go2rtc exec, echo, and expr sources by default. They can be enabled via an environment variable. A separate configuration for this for HA addon users will come in a later beta.
Fixes and improvements to authentication & user management (wait for config before evaluating route access, JWT/token behavior fixes, API auth improvements, password update rules including allowing admin password change without the old password, default unspecified username to viewer)
Role-based access control race-condition fix for custom-role users after login
Mitigate TensorFlow/TFLite atexit crash and logging noise (pre-import on main thread, suppress TensorFlow/TFLite logging and delegate-creation messages, apply base log-level suppressions for noisy third-party libraries)
Batch queries when searching for state images from recordings (performance improvement)
Custom classification wizard improvements (handle remaining/unclassified images when finishing, reset wizard state after closing with a model, make informational popovers modal, disallow "none" as a classification class name, automatically train object classification models when appropriate, validation and UI fixes including long-text wrapping, button overflow prevention, cursor and cleanup fixes)
Use fallback timeout in Live player when opening media sources (handles slow/no go2rtc stream connections)
UI/UX fixes (ensure Explore calendar respects first-day-of-week setting, timeline/detail stream and live camera view fixes, do not render empty actions menus, add sidebar entries, filter camera groups in camera filters to only include groups with at least one allowed camera)
Hardware acceleration and FFmpeg/encoder tweaks (birdseye restream CUDA hwaccel argument fix, NVENC/FFmpeg hwaccel adjustments, handle unknown SoCs for RKNN converter)
Embeddings / GenAI / ONNX fixes (ensure GenAI client exists, fix GenAI callbacks in MQTT, ONNX embedding updates, fix concurrency issue with semantic search when using JinaV2)
Error handling and robustness improvements (catch JSONDecodeError, catch OSError when deleting classification images, ensure no NameError during processing)
Platform / build notes (AVX instruction requirement callout for classification)
Minor frontend fixes (Statusbar GPU key fix, other component cleanups)
Fix thumbnail images remaining in memory indefinitely for GenAI object descriptions in certain circumstances
Review summary report should generate in the language specified by the user
Review / object MQTT handling fix
Documentation updates

Full changelog

Major Changes for 0.17.0

Breaking Changes

There are several breaking changes in this release, Frigate will attempt to update the configuration automatically. In some cases manual changes may be required. It is always recommended to back up your current config and database before upgrading:

Simply copy your current config file to a new location
Stop Frigate and make a copy of the frigate.db file

GenAI now supports reviews and object descriptions. As a result, the global genai config now only configures the provider. Other fields have moved under objects -> genai. See the new GenAI documentation.
Recordings retention is now fully tiered. This means that record -> continuous and record -> motion are separate config fields. See the examples in the documentation.
Some of the LPR models have been updated, and most users should manually switch to the small model, which performs well on both CPU and GPU. The large model is the same as 0.16's and is not as accurate as the upgraded small model in 0.17. Use large only if you live in a region with multi-line plates and you are having issues detecting text on them with the small model.
strftime_fmt was deprecated in 0.16, and should now be fully removed from the config in 0.17. Date/time formatting is based on the language selected in the UI.
The auto detection logic for camera resolution has changed. Some cameras fail to correctly advertise their resolution, and in previous versions, a default value was assumed that was not always correct. You may need to explicitly define detect resolution width and height for cameras in your config if Frigate hangs on startup.
The exec, expr, and echo sources for go2rtc are now removed by default to reduce the security risk if an attacker has access to the configuration. This can be disabled using an environment variable GO2RTC_ALLOW_ARBITRARY_EXEC A separate configuration for this for HA addon users will come in a later beta. See the documentation.

New Features

Frigate 0.17 introduces several major new features.

Classification Model Training

Frigate 0.17 supports classification models in two separate types: state classification and object classification. These models are trained locally on your machine using ImageNet via MobileNetV2.

State Classification

State classification allows you to choose a certain region of camera(s) with multiple states, and train on images showing these states. For example, you could create a state classification model to determine if a gate is currently open or closed.

See the documentation.

Object Classification

Object classification allows you to choose an object type, like dog, and classify specific dogs. For example, you can train the model to classify your dog Fido and add a sub label, while not labeling unknown dogs. Another example would be classifying if a person in a construction site is wearing a helmet or not.

See the documentation.

Custom Viewer Roles

Frigate 0.17 now has the ability to create additional viewer user roles to limit access to specific cameras. Users with the admin role can create a uniquely named role from the UI (or auth --> roles in the config) and assign at least one camera to it. Users assigned to the new role will have:

Guarded API access
Limited frontend access, following what the viewer role has access to (Live, Review/History, Explore, Exports), but only to the assigned cameras

See the documentation.

Review Item Summary with GenAI

Frigate 0.17 supports using GenAI to summarize review items. Unlike object descriptions which add a searchable description, review summaries have a structured output that instruct the AI provider to generate a title, description, and classify the activity as dangerous, suspicious, or normal.

This information is displayed in the UI automatically making it easier to see when activity requires further review and easier to understand what is happening during a particular video segment.

See the documentation.

Semantic Search Triggers

Triggers utilize Semantic Search to automate actions when a tracked object matches a specified image or description. Triggers can be configured so that Frigate executes a specific actions when a tracked object's image or description matches a predefined image or text, based on a similarity threshold. Triggers are managed per camera and can be configured via the Frigate UI in the Settings page under the Triggers tab.

See the documentation.

Object Detector Improvements

Frigate 0.17 brings performance increases for many detectors as well as support for new object detection hardware.

Nvidia GPU Performance

Support for Nvidia GPUs has been enhanced by implementing CUDA Graphs. CUDA Graphs work to reduce the involvement o...

Assets 2

67 Join discussion

14 Dec 14:01

blakeblackshear

v0.17.0-beta1

430cebe

0.17.0 Beta 1 Pre-release

Pre-release

Beta docs: https://deploy-preview-19787--frigate-docs.netlify.app/

Images

Major Changes for 0.17.0

Breaking Changes

Simply copy your current config file to a new location
Stop Frigate and make a copy of the frigate.db file

GenAI now supports reviews and object descriptions. As a result, the global genai config now only configures the provider. Other fields have moved under objects -> genai. See the new GenAI documentation.
Recordings retention is now fully tiered. This means that record -> continuous and record -> motion are separate config fields. See the examples in the documentation.
Some of the LPR models have been updated, and most users should manually switch to the small model, which performs well on both CPU and GPU. The large model is the same as 0.16's and is not as accurate as the upgraded small model in 0.17. Use large only if you live in a region with multi-line plates and you are having issues detecting text on them with the small model.
strftime_fmt was deprecated in 0.16, and should now be fully removed from the config in 0.17. Date/time formatting is based on the language selected in the UI.

New Features

Frigate 0.17 introduces several major new features.

Classification Model Training

State Classification

See the documentation.

Object Classification

See the documentation.

Custom Viewer Roles

Guarded API access
Limited frontend access, following what the viewer role has access to (Live, Review/History, Explore, Exports), but only to the assigned cameras

See the documentation.

Review Item Summary with GenAI

This information is displayed in the UI automatically making it easier to see when activity requires further review and easier to understand what is happening during a particular video segment.

See the documentation.

Semantic Search Triggers

See the documentation.

Object Detector Improvements

Frigate 0.17 brings performance increases for many detectors as well as support for new object detection hardware.

Nvidia GPU Performance

Support for Nvidia GPUs has been enhanced by implementing CUDA Graphs. CUDA Graphs work to reduce the involvement of the CPU for each inference, leading to faster inference times and lower CPU usage. CUDA graphs do have some limitations based on the complexity of the model, which means that YOLO-NAS, Semantic Search, and LPR models are not accelerated with CUDA Graphs. They will still continue to run on GPU as they did before.

Intel OpenVINO

Frigate 0.17 supports running models on Intel NPUs, for many models performance on NPU is similar to GPU but more efficient, leaving room to run more enrichment features on the GPU.

OpenVINO has also had many optimizations put in place to reduce memory and CPU utilization for object detection.

RKNN

Frigate 0.17 brings several improvements to RKNN platform including:

Automatic Model Conversion: automatically convert ONNX models to RKNN format. This allows Frigate+ and other models to be seamlessly configured and converted on startup.
Accelerated Enrichment Support: convert and run Semantic Search and Face Recognition models using the NPU. This greatly enhances performance while maintaining high accuracy with large model sizes.

Apple Silicon

Frigate 0.17 supports running object detection on Apple Silicon NPU. This is provided through the Apple Silicon Detector which runs on the host and connects via IPC proxy to Frigate, providing fast and efficient inferences when run within the same Apple device.

See the documentation.

YOLOv9 on Google Coral

Frigate 0.17 supports running a quantized version of YOLOv9 on Coral devices, bringing improved accuracy over the default mobiledet model. Note that due to hardware limitations, only a subset of the objects on the standard COCO labelmap is included. YOLOv9 Frigate+ models are not supported on Coral at this time.

See the documentation.

New Community Supported Detectors

Frigate 0.17 has community support for several new object detectors:

MemryX: MemryX MX3 M.2 module. Documentation
Degirum SDK: a proxy for inference with a variety of models. Documentation
Synaptics: Synaptics SL1680 NPU. Documentation

Frontend Improvements

In addition to supporting the new features, the frontend has many improvements.

Detail Stream

History view in 0.17 supports an additional view mode, Detail. This mode shows a card for each review item, and expanding a card reveals all tracked objects and their lifecycle events. Selecting any lifecycle event seeks the video to that exact timestamp. You can also overlay a tracked object's path on the video to help with debugging.

Redesigned Tracked Object Details pane

The Tracked Object Details pane in Explore has been redesigned to streamline the layout and consolidate related information. The Object Lifecycle tab is now the Tracking Details tab, which displays video overlays of the tracked object instead of static images, giving a clearer and more intuitive view of its activity.

Revamped Settings

Frigate 0.17 ha...

Assets 2

669 Join discussion

06 Dec 18:15

blakeblackshear

v0.16.3

9034454

0.16.3 Release Latest

Latest

Images

Security Advisory

A security vulnerability was reported and addressed with this release.

Broken access control viewer user can delete admin and other users account

Exploiting this vulnerability requires authenticated access to Frigate.

Improvements

Improve recording playback efficiency
Fix certificate generation crash when using webpush notifications

What's Changed

version bump in docs by @hawkeye217 in #20501
Fix YOLOv9 export script by @NickM-27 in #20514
Fix model exports by @NickM-27 in #20540
Update Azure OpenAI genai docs by @hawkeye217 in #20549
Improve Reolink Camera Documentation by @NickM-27 in #20605
Update PWA requirements and add usage section by @NickM-27 in #20562
Update coral docs / links by @NickM-27 in #20674
Update recommended hardware to list more models by @NickM-27 in #20777
docs: remove webrtc not support H.265 tips by @ZhaiSoul in #20769
Update hailo installation instructions by @NickM-27 in #20847
Fix RF-DETR docs heading by @hawkeye217 in #20963
add comment about unifi g5 and newer cams by @h-leth in #21003
Fixes by @hawkeye217 in #21061
Version bump in docs by @hawkeye217 in #21111
Add Axis Q-6155E camera configuration details by @munit85 in #21105
Pin onnx in rfdetr model generation command by @hawkeye217 in #21127
Pin cryptography package version by @hawkeye217 in #21126
Fix jetson build by @NickM-27 in #21173

New Contributors

@h-leth made their first contribution in #21003
@munit85 made their first contribution in #21105

Full Changelog: v0.16.2...v0.16.3

Contributors

ZhaiSoul, NickM-27, and 3 other contributors

Assets 2

15 Oct 11:30

blakeblackshear

v0.16.2

4d58206

0.16.2 Release

Images

Security Advisory

A security vulnerability was reported and addressed with this release.

Arbitrary File Read via Export Thumbnail "image_path" parameter

Exploiting this vulnerability requires authenticated access to Frigate.

What's Changed

Fix best thumbnail endpoint by @hawkeye217 in #19930
Fix model selection type in Frigate+ settings pane by @hawkeye217 in #19952
Fixes by @NickM-27 in #19984
Improve live loading by not blocking UI by @NickM-27 in #19998
feat: add github mirror download endpoint by @ZhaiSoul in #20007
add robots.txt by @ZhaiSoul in #20093
Improve rf-detr export by @NickM-27 in #20485
Ensure that a user must provide an image in an expected location by @NickM-27 in #20491

New Contributors

@laviddichterman made their first contribution in #19951
@AmirHoseinOmidi made their first contribution in #20306
@mpking828 made their first contribution in #20343

Full Changelog: v0.16.1...v0.16.2

Contributors

laviddichterman, ZhaiSoul, and 4 other contributors

Assets 2

04 Sep 11:07

blakeblackshear

v0.16.1

e664cb2

0.16.1 Release

This is a maintenance release for Frigate 0.16 that includes bugfixes and minor changes.

Images

What's Changed

Frontend Changes

Improve recording playback startup time on slow connections (#19503)
Fix export dialog overflowing due to i18n time lengths (#19736)
Ensure audio support is correctly queried from go2rtc streams (#19708)
Allow users with the viewer role to register for webpush notifications (#19640)
Video dimension layout fix for Chrome (#19636, #19805)
Remind users to restart after enabling face recognition (#19601)
Always allow sorting by score in Explore page (#19501)
Fix recognized license plate filter being incorrectly hidden in Explore filters in certain circumstances (#19491)
Fix unscrollable languages menu on small mobile devices (#19797)
Fix label in streaming stats (#19874)

Backend Changes

Fix i965 hwaccel driver for older Intel iGPUs (#19590)
Return correct mime type for some image endpoints (#19708)
Cleanup vod clip handling and add padding arg (#19813)
Catch invalid keys in genai prompts (#19657)
Run autotracking setup method in asyncio coroutine (#19614)
Fix autotracking calibration crash when zooming is disabled (#19776)
Fix autotracking hang for some cameras (#19873, #19879)
Deprecate include_thumbnails field in events API (#19584)
Ensure bird classification still runs if no other enrichments are enabled (#19576)
Latest image API endpoint fixes (#19573 and #19555)
Fix percentage display in recording emergency cleanup logs (#19525)
Support input-dtype config for Frigate+ models (#19799)
Set lower bound on retry_interval (#19883)

Assets 2

16 Aug 12:30

blakeblackshear

v0.16.0

c2f8de9

0.16.0 Release

Images

Changes since RC 4

None

Major Changes for 0.16.0

Breaking Changes

Simply copy your current config file to a new location
Stop Frigate and make a copy of the frigate.db file

The TensorRT detector has been removed for Nvidia GPUs, the ONNX detector should be used instead. Nvidia recently pushed newer packages and dropped compatibility for older versions of TensorRT. Supporting native TensorRT detector directly has shown to be a large burden, making supporting newer hardware more difficult and increasing the difficulty in updating ONNX runtime. The decision has been made to remove the TensorRT detector which will allow us to focus on supporting more models and bringing more features to the Nvidia platform through ONNX runtime. Users who were running the TensorRT detector will need to update their configuration to use the ONNX detector, and download a new model. YOLO-NAS or YOLOv9 are recommended as they provide superior performance and accuracy compared to the models supported by TensorRT. Documentation
The go2rtc audio handling has been updated to accept any audio codec instead of only AAC. This makes it easier to get audio from cameras that use PCMA/U audio. However, if the recording ffmpeg preset preset-record-generic-audio-copy is currently used it likely will need to be changed to preset-record-generic-audio-aac to continue to support audio in recordings.
```
ffmpeg:
  ...
  output_args: 
    record: preset-record-generic-audio-aac
```
Object detection is now disabled by default. A config migrator will update your config globally with enabled: true for detect:. If you use Frigate with a read-only config, you may need to edit your own config to enable object detection manually.
```
detect:
  enabled: true
```
Frigate will no longer fall back to look for a ffmpeg binary in PATH. If you are extending the Frigate image or somehow injecting a custom ffmpeg binary in a place like /usr/local/bin, please read the custom ffmpeg build guide.
Home Assistant Add-on users will need to be aware of several important breaking changes. See the Important Notes for Home Assistant Add-on Users section below.
The base image for Frigate has been updated to Debian Bookworm. The Jetpack 4 (jp4) and Jetpack 5 (jp5) images have been removed due to not being compatible with the new base image. A new Jetson Jetpack 6 (jp6) build has been contributed by a community member.
The camera group icon pack has been updated. Some of your camera group icons may need to be re-selected.
The Hailo driver and firmware has been updated to 4.21.0. The standard Docker image now provides Hailo support. If you are using Home Assistant Operating System, version 16.0 is required. Other users will need to update the Hailo driver and firmware on their system using the user installation script.
The ROCm migraphx detector has been removed: Instead of separately supporting ROCm via MigraphX and via ONNX, the ROCm (rocm) detector has been removed, and the onnx detector should be used instead. This change should only require changing the type: rocm to type: onnx in detectors
An additional parameter has been added to movement_weights for autotracking users. A config migrator will add the new value so that Frigate starts, but if this fails or you are using relative zooming with autotracking, you should remove movement_weights from your config and recalibrate.
Proxy header mapping for user roles is now required or setting the default_role to admin is required to achieve the same functionality as previous versions. This version implements a Viewer role, which is now the default for authenticated users without a role header map defined in the config. If you are using a proxy, you must include admin in your mapped role header in order for admin users to retain their elevated privileges. See the documentation.
The strftime_fmt config option is now deprecated and will be removed in a future version. The UI will now define time formats based on your language and locale.
Some API routes are now protected and require credentials or changes. With the introduction of the Viewer role and the ability to mark review items as reviewed separately per-user, changes may be needed to any services making calls to Frigate's API.
Some users may need to adjust the tls_insecure onvif config if ONVIF PTZ controls were previously working in past versions but fail to work in 0.16. The ONVIF package was upgraded for 0.16, and several users have reported that setting tls_insecure: false fixed their issues.

Important Notes for Home Assistant Add-on Users

For Home Assistant add-on users only

The Frigate config directory for add-on users was changed from the Home Assistant config directory to the add-on config directory. With this change, the Frigate config file and database are now part of the add-on backup taken by Home Assistant, no longer needing to backup them manually. Note that only backups after 0.16 will benefit of this change. If you took a backup of Frigate 0.15 when updating to 0.16, the Frigate config file and database will not be part of it. In that case, you should still backup them manually before updating to Frigate 0.16.

Automatic migration

Upon the first Frigate 0.16 initialization, all Frigate-related files will be moved from the Home Assistant config directory to the Frigate add-on config directory. This will mean these files will no longer be accessible through your Home Assistant /config directory. They will now be found in the /addon_configs directory, which can also be accessed through the Terminal or VS Code add-ons.

In the new directory, the Frigate config file will be named config.yml rather than frigate.yml. This aligns with people running Frigate through other means. If you want to move back to Frigate 0.15 and did not backup the config file, you can copy the config.yml from the /addons_config back to /config/frigate.yml before restoring the add-on. Consider copying back other files as well, like custom model, custom ffmpeg, and custom go2rtc. Note that the database from 0.16 is not compatible with Frigate 0.15 anyway and thus should not be copied back. To restore the database, you must restore it from a manual backup.

Moving the config directory between add-on variants

The /addon_configs directory contains the config directories of all add-ons. You can move files from one addon config directory to another, like from Frigate Beta to the stable one, or vice-versa. Make sure you do that before uninstalling the add-on, unless you do not select the Also permanently delete this addon's data when uninstalling it.

Breaking change

If your Frigate config file refers to any arbitrary files in the Home Assistant config dir, like echo:/config/ scripts, their references must be manually updated to from /config/ to /homeassistant/.

New Features

Frigate 0.16 introduces several major new features.

Face Recognition

Frigate utilizes face detection and recognition to apply sub-labels to known faces. Recognized faces can be included in built-in notifications and integrated into third-party notifications.

Frigate+ is not required to use Face Recognition. However, using a Frigate+ model (or a custom model that detects face natively) enhances efficiency by eliminating the need for separate face detection per person.

The accuracy of face recognition heavily depends on the quality of the images used for training. It is highly recommended to follow the training guide to achieve the best results.

Refer to the Face Recognition documentation for more.

License Plate Recognition (LPR)

Frigate detects and recognizes license plates, saving recognized plates to tracked car objects. Known license plates can also be assigned custom labels (e.g., AB12345 --> Bob's Car).

Frigate+ is not required for License Plate Recognition. However, using a Frigate+ model (or a custom model that detects license_plate natively) improves efficiency by eliminati...

Assets 2

78 Join discussion

12 Jul 01:14

blakeblackshear

v0.15.2

3bda638

0.15.2 Release

This is a maintenance release for Frigate 0.15 that includes fixes for HAOS 16 "too many open files" errors.

What's Changed

Set ulimit by @NickM-27 in #19086
Fix ulimit setting by @NickM-27 in #19087
Fix ulimit script by @hawkeye217 in #19095
Set ulimit with Python by @hawkeye217 in #19105

Images

Full Changelog: v0.15.1...v0.15.2

Contributors

NickM-27 and hawkeye217

Assets 2

15 Apr 01:30

blakeblackshear

v0.15.1

6cb5cfb

0.15.1 Release

This is a maintenance release for Frigate 0.15 that includes two bugfixes.

What's Changed

Fix crash in get_current_frame() (#16594)
Fix trt model prepare script (#17661)

Images

Assets 2

08 Feb 18:45

blakeblackshear

v0.15.0

cea210d

0.15.0 Release

Images

Changes since RC2

None

Major Changes for 0.15.0

Breaking Changes

There are several breaking changes in this release, Frigate will attempt to update the configuration automatically. In some cases manual changes may be required. It is always recommended to back up your current config and database before upgrading:

Simply copy your current config file to a new location
Stop Frigate and make a copy of the frigate.db file

Rewritten SHM system
- The SHM system has been rewritten to be more stable. This means that a large % of shared memory (SHM) will be used by default but the total usage will not fluctuate, and Frigate will not fail from SHM overflowing. However, this also means that having the correct SHM size for your cameras, detector, and hardware is more important to having a smooth experience with Frigate. A new calculation is done on startup, so if your shm_size is too low then a warning will be printed in the log stating that it needs to be increased.
Recording configuration changed
- The record config has been refactored to allow for direct control of how long alerts and detections are retained. These values will be automatically populated from your current config, but you may want to adjust the values that are set after updating. See the updated docs here and ensure your updated config retains the footage you want it to.
ffmpeg updates
- ffmpeg has been updated to version 7. Some Intel users may need to change what hwaccel preset they are using (preset-vaapi may need to now be preset-intel-qsv-h264 or preset-intel-qsv-h265) if camera feeds are not functioning correctly after upgrading. This may need to be adjusted on a per-camera basis. If a qsv preset is not working properly, you may still need to use a preset-vaapi or revert to the previous ffmpeg version as described below.
- The new version provides many improvements, but the old version can still be run if needed by specifying path: "5.0" in your ffmpeg: config entry. For example:
```
ffmpeg:
  path: "5.0"
```
- ffmpeg presets have been updated. It is always recommended to use ffmpeg presets rather than manual arguments.
- ffmpeg is no longer part of $PATH. In most cases this is handled automatically.
  - go2rtc exec streams will need to add the full path for ffmpeg which in most cases will be /usr/lib/ffmpeg/7.0/bin/ffmpeg.
  - If the go2rtc config has bin: ffmpeg defined, it needs to be removed.
The nested model config under a detector has been simplified to just model_path as a string value. This change will be handled automatically through config migration.
Submitting to Frigate+
- The Frigate+ view in the UI has been replaced by the new Explore view. See the notes below.

Explore

The new Explore pane in Frigate 0.15 makes it easy to explore every object tracked by Frigate. It offers a variety of filters and supports keyword and phrase-based text search, searching for similar images, and searching through descriptive text generated by AI models.

The default Explore pane shows a summary of your most recent tracked objects organized by label. Clicking the small arrow icon at the end of the list will bring you to an infinitely scrolling grid view. The grid view can also be set as the default by changing the view type from the Settings button in the top right corner of the pane.

The Explore pane also serves as the new way to submit images to Frigate+. Filters can be applied to only display tracked objects with snapshots that have not been submitted to Frigate+. The left/right arrow keys on the keyboard allow quick navigation between tracked object snapshots when looking at the Tracked Object Details pane from the grid view.

AI/ML Search

Frigate 0.15 introduces two powerful search features: Semantic Search and GenAI Search. Semantic Search can be enabled on its own, while GenAI Search works in addition to Semantic Search.

Semantic Search

Semantic Search uses a CLIP model to generate embeddings (numerical representations of images) for the thumbnails of your tracked objects, enabling searches based on text descriptions or visual similarity. This is all done locally.

For instance, if Frigate detects and tracks a car, you can use similarity search to see other instances where Frigate detected and tracked that same car. You can also quickly search your tracked objects using an "image caption" approach. Searching for "red car driving on a residential street" or "person in a blue shirt walking on the sidewalk at dawn" or even "a person wearing a black t-shirt with the word 'SPORT' on it" will produce some stunning results.

Semantic Search works by running an AI model locally on your system. Small or underpowered systems like a Raspberry Pi will not run Semantic Search reliably or at all. A dedicated GPU and 16GB of RAM is recommended for best performance.

See the Semantic Search docs for system requirements, setup instructions, and usage tips.

Generative AI

GenAI Search employs generative AI models to create descriptive text for the thumbnails of your tracked objects, which are stored in the Frigate database to enhance future searches. Supported providers include Google Gemini, Ollama, and OpenAI, so you can choose whether you want to send data to the cloud or use a locally hosted provider.

See the GenAI docs for setup instructions and use case suggestions.

Improved Tools for Debugging

Review Item Details Pane

A new Review Item Details pane can be viewed by clicking / tapping on the gray chip on a review item in the Review pane. This shows more information about the review item as well as thumbnails or snapshots for individual objects (if enabled). The pane also provides links to share the review item, download it, submit images to Frigate+, view object lifecycles, and more.

Object Lifecycle Pane

The Recordings Timeline from Frigate 0.13 has been improved upon and returns to 0.15 as the Object Lifecycle, viewable in the Review Details pane as well as the new Explore page. The new pane shows the significant moments during the object's lifecycle: when it was first seen, when it entered a zone, became stationary, etc. It also provides information about the object's area and size ratio to assist in configuring Frigate to tune out false positives.

Native Notifications

Frigate now supports notifications using the WebPush protocol. This allows Frigate to deliver notifications to devices that have registered to receive notifications in the Frigate settings, delivering them in a timely and secure manner. Currently, notifications will be delivered for all review items marked as alerts. More options for native notifications will be supported in the future.

See the notifications docs.

New Object Detectors

ONNX

ONNX is an open model standard which allows for a single model format that can run on different types of GPUs. The default, tensorrt, and rocm Frigate build variants include GPU support for efficient object detection via ONNX models, simplifying configuration and support more models. There are no default included ONNX models for object detection.

AMD MiGraphX

Support has been added for AMD GPUs via ROCm and MiGraphX. Currently there is no default included model for this detector.

Hailo-8

Support has been added for the Hailo8 and Hailo-8L hardware for object detection on both arm64 and amd64 platforms.

Other UI Changes

Rework the Storage metrics pane to be more clear how much storage each camera is using
Camera stream info button has been added to the camera metrics page allowing easy access to see camera codec information
Add button for downloading the full set of logs
The time...

Assets 2

88 Join discussion

29 Aug 22:32

blakeblackshear

v0.14.1

f4f3cfa

0.14.1 Release

Images

This is a maintenance release for Frigate 0.14 that includes bugfixes as well as performance and stability improvements.

What's Changed

Fix auth when serving Frigate at a subpath (#12815)
Fix iOS export buttons (#12755)
Ensure review cameras are sorted by config UI order if specified (#12789)
Fix default model for rknn detector (#12807)
Add button to mark alert as reviewed in filmstrip (#12878)
Add confirmation dialog before deleting review items (#12950)
Fix manual events (#12951)
Add pan/pinch/zoom capability on Frigate+ view snapshots (#12953)
Hide recording button when disabled in config (#12997)
Recordings view fixes (#13005)
Fix missing last hour preview in recordings view (#13027)
Fix export name and NaN showing in Frigate+ slider (#13030)
Prevent error when github sends bad data (#13077)
Add shortcut key "r" to mark items as reviewed in Review pane (#13087)
Fix Frigate+ view from inadvertently reloading (#13160)
Fix live player from getting stuck while loading streams and from lagging behind in poor network conditions (#13143)
Ensure previews that are generated late are not falsely excluded (#13193)
Ensure only enabled birdseye cameras are considered active (#13194)
UI fixes (#13246)
Fix cgroup detecting 0 cpu cores (#13271)
Add tooltip for icons in review event list (#13334)
Don't allow periods in zone or camera group names (#13400)
Add portal to the live player tooltip (#13389)

Assets 2

Uh oh!

Releases: blakeblackshear/frigate

0.17.0 Beta 2

Images

Changes since Beta 1

Major Changes for 0.17.0

Breaking Changes

New Features

Classification Model Training

State Classification

Object Classification

Custom Viewer Roles

Review Item Summary with GenAI

Semantic Search Triggers

Object Detector Improvements

Nvidia GPU Performance

Uh oh!

0.17.0 Beta 1

Images

Major Changes for 0.17.0

Breaking Changes

New Features

Classification Model Training

State Classification

Object Classification

Custom Viewer Roles

Review Item Summary with GenAI

Semantic Search Triggers

Object Detector Improvements

Nvidia GPU Performance

Intel OpenVINO

RKNN

Apple Silicon

YOLOv9 on Google Coral

New Community Supported Detectors

Frontend Improvements

Detail Stream

Redesigned Tracked Object Details pane

Revamped Settings

Uh oh!

0.16.3 Release

Images

Security Advisory

Improvements

What's Changed

New Contributors

Contributors

Uh oh!

0.16.2 Release

Images

Security Advisory

What's Changed

New Contributors

Contributors

Uh oh!

0.16.1 Release

Images

What's Changed

Frontend Changes

Backend Changes

Uh oh!

0.16.0 Release

Images

Changes since RC 4

Major Changes for 0.16.0

Breaking Changes

Important Notes for Home Assistant Add-on Users

New Features

Face Recognition

License Plate Recognition (LPR)

Uh oh!

0.15.2 Release

What's Changed

Images

Contributors

Uh oh!

0.15.1 Release

What's Changed

Images

Uh oh!