Skip to content

Releases: jasonet/Playa

Release list

Playa 0.3.4 Community Preview (Apple Silicon)

Playa 0.3.4 Community Preview (Apple Silicon) Latest
Latest

Choose a tag to compare

View all tags
@jasonet jasonet released this 02 Sep 07:39
v0.3.4
ba58d78
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Playa 0.3.4 Community Preview

Playa 0.3.4 is the first downloadable Apple Silicon Community Preview. It keeps the complete embedded MLX and MLX-VLM Python runtime for local/offline inference after required model files have been downloaded, while deterministic Release pruning keeps the compressed DMG below 300 MiB.

Download

  • Playa-0.3.4-macos-arm64-unnotarized.dmg
  • Apple Silicon (arm64) only
  • macOS 26 or newer
  • Size: 306,671,725 bytes (292.46 MiB)

SHA-256

09209a3e28dd68dafa2c8311407a122cd55a130b0e1d6a8c902b04ba6a82c1f1  Playa-0.3.4-macos-arm64-unnotarized.dmg

Verify after downloading:

shasum -a 256 Playa-0.3.4-macos-arm64-unnotarized.dmg

The output must exactly match the checksum above. A matching .sha256 file is also attached to this release.

Important security notice

This Community Preview is ad-hoc signed and not notarized by Apple. The ad-hoc signature provides code-integrity verification, but it does not establish a trusted publisher identity. Gatekeeper may block the first launch.

Only download this build from the official Playa GitHub release and verify its SHA-256 checksum before opening it.

First launch / Gatekeeper

  1. Open the DMG and drag Playa.app to /Applications.
  2. In Finder, open /Applications.
  3. Control-click Playa and choose Open.
  4. Confirm Open in the dialog.
  5. If macOS still blocks the app, go to System Settings → Privacy & Security, find the Playa warning, and choose Open Anyway.

A future production build signed with a Developer ID Application certificate and accepted by Apple notarization will support the standard macOS launch flow.

Highlights

  • Complete embedded MLX, MLX-VLM, Transformers, scientific Python, image, video, and audio runtime
  • Deterministic Release pruning of bytecode caches, test suites, pip/ensurepip, IDLE, Tk demos, and development headers
  • One-click local MLX and MLX-VLM model serving
  • Local model discovery, filtering, downloads, history, launch, and removal
  • EasyCLIProxyAPI provider configuration and model discovery
  • Classic streaming chat, Fx Agent, OpenComputer execution, and conversational image workflows

Validation performed

  • Mounted-DMG checksum verification
  • Version 0.3.4 and arm64-only architecture verification
  • Strict deep code-sign verification
  • Embedded native Python dependency imports
  • Fx gateway and generated-image bridge regression checks
  • Server startup, /metrics, and clean shutdown lifecycle
Assets 4
Loading

Playa 0.3.3

Playa 0.3.3

Choose a tag to compare

View all tags
@jasonet jasonet released this 01 Sep 14:59
v0.3.3
f782d6d

Playa 0.3.3

Playa's first public source release introduces a native macOS workspace for local and provider-backed AI.

Highlights

  • One-click local MLX and MLX-VLM model serving on Apple Silicon
  • Local model catalog designed for hundreds of models and TB-scale storage
  • Automatic model tags, filtering, discovery, download management, and size sorting
  • EasyCLIProxyAPI configuration with OAuth-compatible upstream providers, model discovery, and local caching
  • Classic streaming chat and multimodal Fx Agent sessions
  • Local-only and optional OpenComputer-enhanced agent execution
  • Conversational image generation and editing through supported Codex and Gemini image models
  • Inline image results, local chat history, performance analytics, endpoint inspection, and server logs

Source build notice

This release publishes source code. Build requirements and setup instructions are documented in README.md. Provider credentials, Apple signing identities, certificates, notarization credentials, and private update keys are intentionally excluded.

Loading