> ## Documentation Index
> Fetch the complete documentation index at: /llms.txt
> Use this file to discover all available pages before exploring further.

# AI Runner Changelog

> Release history for the Livepeer AI Runner, the inference runtime for batch and real-time AI pipelines on the network.

> [!TIP]
>
> This page is an automated workflow.
> Subscribe to this changelog's [RSS Feed](/v2/resources/changelog/ai-runner/rss.xml)

Track releases for [AI Runner](https://github.com/livepeer/ai-runner/releases) on GitHub.

**v0.14.1**

## v0.14.1

*AI Summary*

## What's Changed

This release mainly wraps up the support for custom pipelines with docs and final fixes on metadata routes used by the Orchestrator.

The `scope` pipeline has been fully moved to the [`daydreamlive/scope-runner`](https://github.com/daydreamlive/scope-runner) repository and can be developed without docker, using only local dev environments if preferred.

### Commits

* .github: Add proper tags to base image
* live: Remove scope project
* docs: Create guide on writing custom pipelines
* routes: Fix version and hardware metadata routes

***

[View release on GitHub](https://github.com/livepeer/ai-runner/releases/tag/v0.14.1)

**v0.14.0**

## v0.14.0

*AI Summary*

## What's Changed

### 🎬 StreamDiffusion Pipeline – Complete Overhaul

* **Prompt & Seed Blending**: Smooth interpolation between values
* **ControlNet Support**: Depth, canny, and other controlnets with dynamic scale adjustment
* **IPAdapter**: Style transfer from reference images. FaceID version for face-preserving generation
* **SD1.5 + SDXL**: Expanded model architecture support
* **NSFW Detection**: Freepik-based safety checker
* **Dynamic Parameters**: In-place updates for most parameters without requiring pipeline restart (eg t\_index\_list, controlnets, etc)
* **Dynamic Resolution TensorRT**: Single engine with support for different resolutions (still requires a reload)
* **TemporalNet ControlNet**: Optical flow-based conditioning for frame-to-frame temporal consistency
* **Multi-stage pipeline**: Configurable image or latent, pre and post-processing in the pipeline (e.g.: latent feedback, upscaling, skip diffusion altogether)
* **StreamV2V**: Cached attention for even further temporal consistency

### 🔧 Realtime Runtime

* **Extensible runner** – Modular pipeline architecture with plugin pattern (#866)
* **Loading screen** – Visual feedback during pipeline reloads (#739)
* Removed Conda and moved to `uv` (except ComfyUI) (#854)
* Higher performance and lower latency
* Improve error handling and monitoring

### 📦 Batch Pipelines

* **Audio-to-Text**: Added faster-whisper backend (#595)
* **LLM**: Bumped vLLM version (#705)

## New Contributors

* @vcashwin made their first contribution in [https://github.com/livepeer/ai-runner/pull/819](https://github.com/livepeer/ai-runner/pull/819)
* @corey-livepeer made their first contribution in [https://github.com/livepeer/ai-runner/pull/837](https://github.com/livepeer/ai-runner/pull/837)

***

[View release on GitHub](https://github.com/livepeer/ai-runner/releases/tag/v0.14.0)

**v0.13.8**

## v0.13.8

*AI Summary*

### New features & improvements

* Support for audio passthrough and Opus encoding *(real-time)*
* Faster and more reliable startup & teardown *(real-time)*
* Improved logging, metrics, and debugging *(real-time)*
* Zero-copy GPU tensor transfer for lower latency *(real-time)*
* Streamdiffusion pipeline improvements and re-introduction *(real-time)*

### Performance & stability

* Many encoder and decoder fixes *(real-time)*
* Better pipeline cleanup and failover handling *(real-time)*
* Robust support for multi-stream and resolution changes *(real-time)*
* Docker image optimisations and workflow improvements

### Developer & maintenance updates

* Moved ai-worker package to go-livepeer
* Updated LLM pipeline dependencies to improve performance and fix bugs *(batch)*
* Improved CI workflows and build caching
* Enhanced checkpoint and model management

### New contributors

* @Radovenchyk – first contribution \[[#417](https://github.com/livepeer/ai-runner/pull/417)]\([https://github.com/livepeer/ai-runner/pull/417](https://github.com/livepeer/ai-runner/pull/417))
* @livepeer-robot – first contribution \[[#611](https://github.com/livepeer/ai-runner/pull/611)]\([https://github.com/livepeer/ai-runner/pull/611](https://github.com/livepeer/ai-runner/pull/611))

***

[View release on GitHub](https://github.com/livepeer/ai-runner/releases/tag/v0.13.8)

**v0.13.3**

## v0.13.3

*AI Summary*

## What's Changed

* Support for JSON string in defaultImage field for overriding pipeline specific images in the mappings

***

[View release on GitHub](https://github.com/livepeer/ai-runner/releases/tag/v0.13.3)

**v0.13.2**

## v0.13.2

*AI Summary*

## What's Changed

* examples: Add test run for lv2v
* dl\_checkpoints.sh - docker run as a user not the root
* dl\_checkpoints.sh run as root but chown to current user
* refactor(runner): update LLM 'tokens\_used' to usage

***

[View release on GitHub](https://github.com/livepeer/ai-runner/releases/tag/v0.13.2)

**v0.13.1**

## v0.13.1

*AI Summary*

## What's Changed

* feat: update A2T audio conversion
* Remove liveportrait
* feat: better LLM response format
* Fix codegen and some small updates

***

[View release on GitHub](https://github.com/livepeer/ai-runner/releases/tag/v0.13.1)

**v0.13.0**

## v0.13.0

*AI Summary*

This release introduces several new ComfyUI pipelines, including Florence 2, LivePortrait, and Sam2, alongside significant performance optimisations and bug fixes for both the realtime AI stack and non-realtime pipelines like A2T and LLM. These updates enhance processing efficiency, stability, and functionality across the platform.

***

[View release on GitHub](https://github.com/livepeer/ai-runner/releases/tag/v0.13.0)

**v0.12.6**

## v0.12.6

*AI Summary*

## What's Changed

* live/runner: Increase liveportrait restart threshold
* Allow running worker on non-GPU machine
* fix: prevents nvml runtime error
* api: Fix startup of control API
* Add error handling to control message handler task
* runner: add events url to live video to video
* runner: Make control\_url and events\_url optional
* SAM2 video-to-video real-time pipeline
* fix: managed containers return after inference completed

## New Contributors

* @gioelecerati made their first contribution in [https://github.com/livepeer/ai-worker/pull/303](https://github.com/livepeer/ai-worker/pull/303)

***

[View release on GitHub](https://github.com/livepeer/ai-runner/releases/tag/v0.12.6)

**v0.12.5**

## v0.12.5

*AI Summary*

## What's Changed

* feat: add hardware info reporting
* feat(worker): auto pull images if not found

***

[View release on GitHub](https://github.com/livepeer/ai-runner/releases/tag/v0.12.5)

**v0.12.4**

## v0.12.4

*AI Summary*

## What's Changed

* docker: ComfyUI model downloading/caching
* worker,runner: Fix lifecycle management of realtime containers
* test: add b64 functions tests
* worker: Add noop image to map
* Fix control\_url task runner

***

[View release on GitHub](https://github.com/livepeer/ai-runner/releases/tag/v0.12.4)


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.