v0.40.12 adds the LFM3 family of small multimodal models and d1 2B decision models, expanding local AI capabilities on Apple silicon. It also delivers performance improvements and bug fixes without introducing breaking changes.
Hugging Face Transformers v5.21 introduces support for multiple new models (such as MiniCPM-OCR, Flint, UltraVBert, and Gradium/Pocket TTS) and enforces breaking architectural changes including custom tensor parallelism and MoE expert sharding. Developers upgrading to this version should review updated TP plans and custom MoE implementations.
Twenty twenty/v2.51.0. Source highlight: Run `upgrade:2-51:migrate-rich-text-to-tiptap` after upgrading to convert existing rich text records to TipTap format. (#28486)
Open WebUI v0.12.4 introduces major performance enhancements such as skill lazy loading and context compaction alongside granular permission enforcement and extensive UI accessibility upgrades. Users will experience a more streamlined interface with context-aware controls, improved state persistence, and critical security fixes.
Appwrite 1.9.10. Source highlight: Appwrite 1.9.10 is a security release for self-hosted deployments on the 1.9 line. It backports the hardening shipped in 2.6.2 across auth, API keys, databases, realtime, GraphQL, storage, VCS, functions, migrations and messaging, and moves the runtime to a patched Brotli. It keeps Console 8.7.37 fo
Appwrite 2.6.2. Source highlight: Appwrite 2.6.2 is a security release for self-hosted deployments on the 2.6 line. It brings the hardening work merged since 2.6.1 across auth, API keys, databases, realtime, GraphQL, storage, VCS, functions, sites, migrations and messaging. It also moves the runtime to a patched Brotli and the Conso
Biome @biomejs/[email protected]. Source highlight: Added the new nursery rule `noArraySortForMinMax`, which reports sorting an array only to read its smallest or largest value. Use `Math.min()` or `Math.max()` instead.
Ollama v0.40.11 adds speech generation on macOS via the `ollama run` command and an OpenAI-compatible `/v1/audio/speech` endpoint, plus introduces d3 decision models for classification, yes/no answers, and probabilistic scoring. These features expand Ollama's multimodal capabilities and structured reasoning without introducing breaking changes.
Ollama version v0.40.10 introduces support for NVIDIA's Nemotron 3.5 Bolt E4B vision-language model. Developers can now run this 128K-token context model locally with speculative decoding support.
v2.22.1 adds the ability to update payment session client data after creation and seeds the fa-IR locale as the default translation, while delivering numerous bug fixes and UI improvements across the dashboard and core flows. The update is a routine patch with no breaking changes, allowing developers to adopt it safely without migration effort.
PocketBase version v0.40.6 introduces critical security updates, memory usage protections for thumbnail generation, and shifts future update distribution away from GitHub. Developers should update promptly to mitigate Go security vulnerabilities and benefit from dependency upgrades.
PocketBase v0.22.57 introduces a memory-safe limit for image thumbnail generation and updates dependencies to address a security vulnerability related to the eMudhra CA compromise. These changes improve system resilience against excessive memory consumption and enhance overall security posture.
Official Neon product update: Jev on the AI Gateway, automatic top-up for AI Gateway credits, an OAuth provider for Managed Better Auth, and more release - Oct 23, 2026. Read the publisher’s changelog for full details.
Linear 2026-10-22-more-ways-to-run-and-sha. Source highlight: Loops automate product work by responding to changes in Linear and connected tools. You can now trigger them from Slack and GitHub, share them as templates, and customize their reasoning levels.
LocalAI v4.12.0 introduces a redesigned web UI, streaming voice responses, native decision routing, and persistent Qdrant vector stores. The release also implements advanced privacy controls, remote MCP OAuth2 authentication, and live GEM-X motion capture.
Raycast v2.9. Source highlight: This update brings new commands, useful improvements, and plenty of fixes across Raycast. Send Notes directly to AI Chat, restart the frontmost app with a new command, and enjoy a more polished AI Chat experience. Search performance is improved by 50% in Root Search and menus, while Search Screensho
Bun version 1.4.4 has been released with routine maintenance updates and contributions from 39 developers. Users can install or upgrade using standard installation scripts or the built-in upgrade command.
Ollama v0.40.9 brings Flint support to Linux and Windows platforms while improving error handling for streaming OpenAI and Anthropic API endpoints. Additionally, it resolves context length propagation bugs in `ollama launch` and refines thinking configuration logic for web search follow-up requests.
Appwrite 1.9.9. Source highlight: Appwrite 1.9.9 is a precautionary maintenance release for self-hosted deployments on the 1.9 line. It moves the runtime to PHP 8.5.12. It retains Console 8.7.37 for compatibility with the 1.9 server APIs. Upgrading from 1.9.8 keeps the existing migration class and introduces no schema changes.
Appwrite 2.6.1. Source highlight: Appwrite 2.6.1 is a precautionary maintenance release that moves the runtime to PHP 8.5.12. It ships with the same 1.2.97 self-hosted Console as 2.6.0, has no schema changes, and upgrading from 2.6.0 only replaces the images.