Ollama

Track Ollama development. Run large language models locally.

About this developer podcast

Ollama turns GitHub activity into short audio updates.

Episodes cover commits, pull requests, issues, and release work.

Use this page to listen, subscribe, and find related developer podcasts.

This show page is the public hub for the podcast. It has one stable listen URL.

Use it to find RSS feeds, recent episodes, related shows, and older episode links.

If the project is quiet, new episodes may pause until there is useful activity to summarize.

Recent episodes from Ollama

  1. Structured Output Meets Speculative Decoding

    The MLX runner now supports speculative decoding and grammar-constrained output at the same time, closing a performance gap that previously forced a choice between the two. Supporting CI and parser fixes round out a release focused on…

  2. Hardening the MLX Engine and Fixing What Feeds the Scheduler

    The MLX engine got a reliability pass, catching silent failures and adding image, audio, and cached-token reporting, while a cluster of scheduler and memory PRs worked to fix inaccurate VRAM and context-length numbers that drive…

  3. MLX Grows Up, Model Intent Gets Respected

    A cluster of Daniel Hiltgen's PRs matured MLX from an experimental backend into a properly tested, version-pinned component, while a separate fix restored model-authored generation defaults that Ollama's built-in settings had been…

  4. Runner Reliability and a Documentation Cleanup Wave

    The MLX runner picked up context-length enforcement and repeat-loop protection to match the llama.cpp backend's behavior, while a separate contributor pushed through nine documentation fixes in a single morning. Together they show a…

  5. Proxy Fix and a Documentation Cleanup Sweep

    A single security-relevant fix closes a proxy gap in Ollama's file transfer path, while four separate documentation pull requests tidy up grammar and formatting in the README and API docs.

  6. Weekly Recap - Claude Desktop Integration & MLX Engine Maturity

    This week's work centered on two fronts: a major push to make Claude Desktop a first-class integration across macOS and Windows, and a series of correctness fixes hardening the MLX engine for structured output, model loading, and Windows…

  7. Correctness Fixes Under the Hood

    Today's activity centers on quiet correctness fixes — a GGUF parsing bug that silently corrupted tensor alignment, a Vulkan regression that broke model loading on integrated GPUs, and stricter but more forgiving API schema handling —…

  8. Agent Expansion, Desktop Polish, and a Documentation Sweep

    Ollama's agent and desktop surfaces both grew today, with new environment-aware tooling and Windows Claude Desktop support landing alongside a performance fix for structured output on the MLX runner. A separate wave of small…

  9. MLX Engine Grows Up
  10. Desktop App Stability and the Claude Integration Cleanup
  11. MLX Runner Hardens Up, Claude Desktop Gets Real Estate
  12. Claude Desktop Integration Overhaul
  13. Hardening the Runner Scheduler
  14. Weekly Recap - Desktop Apps Grow Up, Inference Gets More Reliable
  15. Visibility Into What the Model Actually Sees
  16. Fixing the Long-Prompt Timeout Trap
  17. Claude Desktop Gets Deep App Integration
  18. Closing the Gaps on Hung Requests and Broken Installs
  19. Closing the Gap on Model Metadata Overhead
  20. Silent Failures Get Loud
  21. Fixing the Small Gaps Between Local and Cloud
  22. Weekly Recap - Model Compatibility and the Push Into Coding Agents
  23. Silent Failures Get Louder
  24. Qwen Gets Serious About Coding Agents
  25. Launch Ecosystem Expands, Parser Edge Cases Get Squashed
  26. Backup Collisions and a Security Fix Converge
  27. Fixing the Defaults That Silently Hurt You
  28. Model Coverage and Tool-Call Reliability
  29. Vision Support Lands in the MLX Runner
  30. Weekly Recap - Vision Runners and Robustness Cleanup
  31. Edge Case Cleanup Across the Stack
  32. Speculative Decoding Gets a New Draft Architecture
  33. Crash Fixes and the Launcher Ecosystem Grows
  34. Silent Data Corruption Fixes
  35. Hardening the Thinking and Tool-Call Pipeline
  36. Closing the Gaps Where Failures Go Silent
  37. Streams That Fail Silently, Fixed
  38. Weekly Recap - Speed, Speculation, and Scheduler Reliability
  39. Scheduler Reliability Overhaul
  40. Speculative Decoding and the Cloud Model Nudge
  41. API Compatibility and Model Correctness Push
  42. Speculative Decoding Gains and a Lint Lockdown
  43. Cleaning Up Concurrency and Cutting Experimental Code
  44. Trust Your Cache, Fix Your Logs
  45. Parsing Bugs and Process Hardening
  46. Weekly Recap - New Model Support & Concurrency Hardening
  47. Closing the Gaps Between Local Config and Cloud Compatibility
  48. MLX Performance Push and a Scheduler Race Fix
  49. A Server Reliability Sweep
  50. Tool Calling Gets a Reliability Pass
  51. Scheduler Race Conditions Take Center Stage
  52. Command Line Consolidation and Hardware Fixes
  53. Thinking Models Leak Control Tokens
  54. Weekly Recap - Hardening the Runtime, One Edge Case at a Time
  55. Closing the Gaps Between API Layers
  56. Cleaning Up the Agent Loop and Widening Model Support
  57. Agent Package Gets a Major Consolidation
  58. Agent Hardening and a Path Traversal Fix
  59. Resource Leaks and GPU Placement Get a Clean-Up Pass
  60. Cleaning Up Memory Planning and the Launch Experience
  61. Fixing What Defaults Got Wrong
  62. Weekly Recap - Model Correctness and Runtime Hardening
  63. Tightening Up the Serving Layer
  64. Thinking Model Output Is Getting a Real Fix
  65. Stability Sweep Across Cloud, GPU, and Agent Tools
  66. Qwen3.5 Gets Untangled, and Small Bugs With Big Blast Radius
  67. Model Behavior Bugs and Blob Security Hardening
  68. Launch Gets Smarter, Model Support Gets Wider
  69. Cleaning Up the Scheduler's Edge Cases
  70. Weekly Recap - Rebuilding the Model Pipeline and Tightening the Guardrails
  71. Truth in Reporting
  72. MLX Create Pipeline Rewrite Lands
  73. Agent Harness Lands, Hardware Support Gets a Cleanup
  74. Gemma 4 Support and Platform Improvements
  75. Weekly Recap - MLX Performance & Path Handling
  76. Memory Management and Multimodal Parsing Fixes
  77. GPU Offloading and Tool Call Fixes
  78. Performance Optimizations and Model Handling Improvements
  79. Infrastructure Updates and Platform Fixes
  80. Multimodal Fixes and Developer Experience Updates
  81. Cache Architecture Overhaul and Data Race Fixes
  82. Developer Tools and Cross-Platform Reliability
  83. Weekly Recap - Integration Expansion & Server Reliability
  84. Audio Support and Infrastructure Refinements
  85. Integration Ecosystem and API Consistency Push
  86. Platform Integration Expansion and API Reliability Fixes
  87. Model Integration and Windows System Improvements
  88. LLaMA Server Integration Hardening
  89. Integration Platform Expansion
  90. Model Integration Updates
  91. Weekly Recap - Infrastructure Modernization
  92. Major Architecture Overhaul Removes CGO Dependencies
  93. MLX Model Display Fixes and Template Parser Cleanup
  94. Weekly Recap - Performance Optimization & Launch System Improvements
  95. DFlash Speculative Decoding Rollback
  96. Model Inventory Refactoring
  97. Startup Performance Optimization
  98. Codex Integration Enhancement
  99. Weekly Recap - MLX Performance & Codex Integration
  100. Release Build Optimization

Episode archive for Ollama

  1. Speculative Decoding and Codex App Updates
  2. MLX Sampler Overhaul and Codex Integration
  3. Vision Model Integration Enhancement
  4. MLX Threading and Claude Image Fixes
  5. Model Transfer Optimization and Test Reliability
  6. Claude Desktop Integration Removed
  7. Launch Command Enhancements
  8. Speed Revolution - MTP Decoding and Smart Caching
  9. Go 1.26 Runtime Update
  10. Weekly Recap - MLX Threading & Model Recommendations
  11. MLX Threading Fixes and Claude App Integration
  12. Model Recommendations and Windows Gateway Fix
  13. Metal GPU Stability and Gemma4 Updates
  14. Launch Experience Improvements and Model Recommendations
  15. Multi-Sequence Batching and New Model Support
  16. Tokenizer Bug Fix for BPE Processing
  17. Weekly Recap - MLX Performance & Launch Integrations
  18. MLX Sampling Performance Enhancement
  19. OpenAI Reasoning Integration
  20. Launch System Improvements and Integration Fixes
  21. Launch System Overhaul and Documentation Updates
  22. MLX Performance Boost and Model Updates
  23. New CLI Integration and Performance Improvements
  24. Weekly Recap - MLX Performance & Launch Integration Expansion
  25. MLX Sampler Improvements
  26. Windows WSL Integration Simplified
  27. Gemma4 Enhancements and Copilot CLI Integration
  28. Hermes Agent Integration and Gemma4 Improvements
  29. Gemma 4 MLX Support and Mixed-Precision Improvements
  30. Weekly Recap - Model Integration and Tooling Enhancements
  31. ROCm 7.2.1 Performance Update
  32. Gemma4 Parser Improvements
  33. Model Updates and Tool Call Fixes
  34. Error Handling and Modelfile Fixes
  35. Weekly Recap - Gemma4 Integration & Audio Support
  36. Performance Lessons and Gemma4 Refinements
  37. Gemma4 Arrives with Audio Magic
  38. Modernizing Codex Configuration
  39. Tokenizer Love and Better Model Support
  40. Legacy Compatibility and Developer Experience Wins
  41. Smoothing the Launch Experience
  42. Fixing the Inconsistencies That Matter
  43. Smart Caching and Better User Experience
  44. VS Code Integration Takes Center Stage
  45. Precision Revolution - New Float Formats and Testing Powerhouse
  46. MLX Performance Breakthrough and Smarter Caching
  47. Nvidia Partnership Takes Center Stage
  48. Bug Squashing Bonanza
  49. The Caching Revolution
  50. Bug Squashing and Launch Improvements
  51. Launch Command Gets a Major Polish
  52. Spring Cleaning and Performance Gains
  53. Thinking Streams and Local Tool Power-ups
  54. Stability First - Error Handling and Performance Fixes
  55. MLX Gets a Major Upgrade and Web Search Goes Live
  56. Simplifying the Sampling Story
  57. Cloud Models Get Smarter & Build Performance Boost
  58. Cloud Integrations Get Some Love
  59. Smarter Constraints and Qwen3.5 Boost
  60. Cloud Integration Drama and AI Model Expansion
  61. Smarter Sampling and Crash Prevention
  62. Building Bridges for Better Model Compatibility
  63. MLX Runner Gets Rock Solid
  64. Tool Calling Gets Smarter
  65. Cleaner Shutdowns and Faster Startups
  66. Qwen 3.5 Architecture Lands with Safety Upgrades
  67. Memory Management Revolution
  68. Nemotron Architecture Lands with Unified Cache Vision
  69. Fixing the WSL Plugin Problem
  70. Smarter UIs and Smoother Onboarding
  71. Tokenizer Consolidation & MLX Library Improvements
  72. Rolling Back and Rolling Forward
  73. Editor Integration Revolution
  74. MLX Display Bug Squashing Day
  75. MLX Runner Gets Major Model Upgrades
  76. MLX Performance Breakthrough and Anthropic Search
  77. MLX Runner Revolution and Documentation Polish
  78. Refactoring Rollercoaster and Developer Experience Wins
  79. Bug Squashing Bonanza
  80. Smooth Onboarding for New Users
  81. Polish and Perfectionism - The Art of Getting the Details Right
  82. Cleaning Up the Config Game
  83. Speed Boost and Model Magic
  84. Memory Magic and Command Makeover
  85. Making Ollama Play Nice with Everyone
  86. The Great Cleanup - Manifests Get Their Own Home
  87. New Model Architecture and Image Generation Fixes
  88. New Model Support and Memory Management Wins
  89. FLUX.2 Image Generation Arrives
  90. Image Generation Goes Native and Parser Cleanup Magic
  91. Dynamic Loading and Experimental Models Take Center Stage
  92. Release Day Rescue Mission

More public developer podcasts

Browse other public Podlog show hubs with crawler-readable RSS and episode links.

  1. Next Js Daily 56 episodes
  2. Headroom Daily 65 episodes
  3. Agent Of Empires Daily 56 episodes
  4. Buzz Transcription 81 episodes
  5. Frigate NVR Updates 199 episodes
  6. Linux Kernel 190 episodes
  7. Homebrew 211 episodes
  8. TypeScript 100 episodes
  9. Ruby on Rails 204 episodes
  10. Redis 179 episodes
  11. Next.js 213 episodes
  12. Vue.js 152 episodes
  13. Rust 183 episodes
  14. Ruby Core Updates 206 episodes
  15. Python 196 episodes
  16. Go 196 episodes
  17. VS Code 214 episodes
  18. Node.js 208 episodes
  19. Django 200 episodes
  20. React Native 157 episodes
  21. PostgreSQL 200 episodes
  22. Kubernetes 209 episodes
  23. LangChain 194 episodes
  24. PyTorch 213 episodes
  25. Openharness Daily 28 episodes
  26. Maestro Daily 55 episodes
  27. Navidrome Daily 74 episodes
  28. Next.js Daily 215 episodes
  29. Pi Mono 95 episodes
  30. OpenClaw 172 episodes
  31. RuView 133 episodes
  32. Shannon 85 episodes
  33. Godot Daily 160 episodes
  34. Agora Next Updates 230 episodes
  35. Rails Daily 214 episodes
  36. Home Assistant Daily 219 episodes
  37. Linux Kernel Daily 195 episodes
  38. React Daily 111 episodes
  39. Jabref Daily 66 episodes
  40. BlocksBeyondTheStars Daily 48 episodes
  41. Jabref Daily 67 episodes
  42. NativeScript iOS Daily 27 episodes
  43. Videorc Daily 53 episodes
  44. iCloud Photos Downloader 27 episodes
  45. Onlook Design Updates 34 episodes
  46. The Algorithm Daily 15 episodes
  47. TailwindCSS 119 episodes
  48. AtomVM Daily 15 episodes
  49. Agora Next Daily 12 episodes
  50. Energy Inc Ace Mate Learning Daily 2 episodes
  51. tiny-gpu Daily 20 episodes
  52. OpenAI Skills 52 episodes
  53. Crush Daily 1 episode
  54. ControlNet Daily 0 episodes
  55. FinceptTerminal Daily 0 episodes