v0.4.4

What's Changed

• fix(converter): move keep_data_uris to convert call and fix lint by @RussellLuo in #165
• feat(mutation): route serial add mutations through coordinator by @gwokhou in #156
• fix(agent): make parallel_tool_calls configurable to fix Amazon Bedrock query/chat (#175) by @KylinMountain in #177
• feat: upgrade pageindex sdk and compile concurrency configurable by @KylinMountain in #174

Full Changelog: v0.4.3...v0.4.4

What's Changed

• fix(converter): move keep_data_uris to convert call and fix lint by @RussellLuo in #165
• feat(mutation): route serial add mutations through coordinator by @gwokhou in #156
• fix(agent): make parallel_tool_calls configurable to fix Amazon Bedrock query/chat (#175) by @KylinMountain in #177
• feat: make indexing and compile concurrency configurable by @KylinMountain in #174

Full Changelog: v0.4.3...v0.4.4

v0.4.4-rc1

What's Changed

• fix(converter): move keep_data_uris to convert call and fix lint by @RussellLuo in #165
• feat(mutation): route serial add mutations through coordinator by @gwokhou in #156
• fix(agent): make parallel_tool_calls configurable to fix Amazon Bedrock query/chat (#175) by @KylinMountain in #177

Full Changelog: v0.4.3...v0.4.4-rc1

v0.4.3

What's Changed

• fix(lint): don't hide same-stem orphans behind qualified links by @jichaowang02-lang in #147
• fix(compiler): strip cache_control for non-Anthropic providers by @Aldominguez12 in #154
• feat: add serial crash-safe mutation recovery by @gwokhou in #142
• perf(mutation): don't snapshot the whole blob store on every add by @KylinMountain in #155
• Agent-first dev harness: AGENTS.md map, CI gate, golden principles, file-size lint by @KylinMountain in #159
• fix(images): disambiguate relative images that share a basename by @KylinMountain in #160
• fix(compiler): guard concept/entity generation against malformed & truncated LLM output by @KylinMountain in #161
• fix(converter): preserve docx embedded images from markitdown by @RussellLuo in #163

New Contributors

• @Aldominguez12 made their first contribution in #154
• @RussellLuo made their first contribution in #163

Full Changelog: v0.4.2...v0.4.3

What's Changed

• fix(lint): don't hide same-stem orphans behind qualified links by @jichaowang02-lang in #147
• fix(compiler): strip cache_control for non-Anthropic providers by @Aldominguez12 in #154
• feat: add serial crash-safe mutation recovery by @gwokhou in #142
• perf(mutation): don't snapshot the whole blob store on every add by @KylinMountain in #155
• Agent-first dev harness: AGENTS.md map, CI gate, golden principles, file-size lint by @KylinMountain in #159
• fix(images): disambiguate relative images that share a basename by @KylinMountain in #160
• fix(compiler): guard concept/entity generation against malformed & truncated LLM output by @KylinMountain in #161
• fix(converter): preserve docx embedded images from markitdown by @RussellLuo in #163

New Contributors

• @Aldominguez12 made their first contribution in #154
• @RussellLuo made their first contribution in #163

Full Changelog: v0.4.2...v0.4.3

v0.4.2

What's Changed

✨ New Features

• Interactive knowledge graph visualization (3D, mind-map, radial) by @KylinMountain in #103
• Configurable LLM request timeout by @KylinMountain in #134
• Global LiteLLM settings via a litellm: config block by @KylinMountain in #138
• Import existing PageIndex Cloud indices via add --from-pageindex-cloud (closes #88) by @KylinMountain in #97
• Google OKF‑conformant wiki frontmatter by @KylinMountain in #102

A glimpse of the new openkb visualize command, showing an interactive 3D knowledge graph.

🐛 Bug Fixes

• Fix resolution of wikilinks whose stem contains a dot (.) by @jichaowang02-lang in #120
• Preserve frontmatter when closing --- has no trailing newline by @jichaowang02-lang in #121

📚 Documentation & Examples

• Multiple README updates and badge improvements by @rejojer (#107 – #119, #124, #126 – #129)
• OpenKB usage guides with real generated artifacts by @KylinMountain in #141 ,#144 and #145

👥 New Contributors

• @jichaowang02-lang made their first contribution in #120

Full Changelog: v0.4.1...v0.4.2

v0.4.2-rc1

What's Changed

🚀 Features

• Interactive knowledge graph visualization (3D, mind-map, radial) via openkb visualize by @KylinMountain in #103
• Google OKF-conformant wiki frontmatter by @KylinMountain in #102
• Configurable LLM request timeout by @KylinMountain in #134
• Pass through global LiteLLM settings via a litellm: config block by @KylinMountain in #138

A glimpse of the new openkb visualize command, showing an interactive 3D knowledge graph.

🐛 Bug Fixes

• Resolve references/wikilinks whose stem contains a dot by @jichaowang02-lang in #120
• Preserve frontmatter when closing "---" has no trailing newline by @jichaowang02-lang in #121

📝 Documentation

• Multiple README updates and badge improvements by @rejojer

New Contributors

• @jichaowang02-lang made their first contribution in #120

Full Changelog: v0.4.1...v0.4.2-rc1

New Contributors

• @jichaowang02-lang made their first contribution in #120

Full Changelog: v0.4.1...v0.4.2-rc1

v0.4.1

✨ Highlights

🎨 New generator: turn your knowledge into a shareable slide deck

openkb deck new effective-agents "Explain how to build effective LLM agents to engineers"
# → output/decks/attention/index.html · open, full-screen, ←/→/F/P to navigate

A single self-contained HTML deck compiled from your wiki — no build, no dependencies. Magazine-style by default (openkb-deck-editorial).

openkb deck new attention "..." --skill deck-swiss-international   # swap visual direction
openkb deck new attention "..." --critique                        # html-critic polishes CSS bugs, a11y, self-containment

Example deck：

🏷️ Entity pages — a new first-class wiki layer

wiki/entities/ now tracks specific named things (people, orgs, places, products), distinct from abstract concepts/. Entity extraction is folded into the existing concept-plan call — no extra LLM request. Ask-the-wiki routes "who/what is X" straight to entity pages. Vocabulary is configurable via entity_types; backfill existing KBs with openkb recompile --all.

---

What's Changed

🚀 Features

• feat(skills): add openkb-deck-neon (dark Aurora Glass) by @KylinMountain in #101
• feat(deck): HTML slide-deck generator + universal skill loader by @KylinMountain in #68
• feat: entity pages — first-class entities/ type (configurable) + recompile backfill by @KylinMountain in #78
• feat: add KB mutation locks and atomic state writes by @gwokhou in #86

🐛 Bug Fixes

• Eliminate duplicate wiki-tool implementations in skill/tools.py by @wutongyuonce in #79
• fix: sync plugin manifest version to v0.3.0 by @moyu12-ae in #82
• fix: harden long PDF page extraction by @gwokhou in #85
• fix(compiler): don't cap concepts-plan at max_tokens=2048 (empty plans on reasoning models) by @cnndabbler in #90
• fix(cli): close litellm async clients after each compile/index (CLOSE-WAIT/FD leak) by @cnndabbler in #91
• refactor(cli,compiler): close litellm async clients inside the compile loop by @KylinMountain in #95
• fix: collision-resistant doc_name — same-stem documents no longer overwrite each other by @KylinMountain in #96
• fix: support ChatGPT-subscription & GitHub Copilot providers (litellm 1.87.2 + extra_headers) by @KylinMountain in #98
• fix(locks): cross-platform file locking — fix Windows fcntl crash by @KylinMountain in #99
• refactor(locks): delegate cross-platform file locking to portalocker by @KylinMountain in #100

New Contributors

• @moyu12-ae made their first contribution in #82
• @gwokhou made their first contribution in #85
• @cnndabbler made their first contribution in #90

Full Changelog: v0.3.0...v0.4.0

v0.4.0

What's Changed

• feat(deck): HTML slide-deck generator + universal skill loader by @KylinMountain in #68
• Eliminate duplicate wiki-tool implementations in skill/tools.py by @wutongyuonce in #79
• feat: entity pages — first-class entities/ type (configurable) + recompile backfill by @KylinMountain in #78
• fix: sync plugin manifest version to v0.3.0 by @moyu12-ae in #82
• fix: harden long PDF page extraction by @gwokhou in #85
• fix(compiler): don't cap concepts-plan at max_tokens=2048 (empty plans on reasoning models) by @cnndabbler in #90
• fix(cli): close litellm async clients after each compile/index (CLOSE-WAIT/FD leak) by @cnndabbler in #91
• refactor(cli,compiler): close litellm async clients inside the compile loop by @KylinMountain in #95
• fix: collision-resistant doc_name — same-stem documents no longer overwrite each other by @KylinMountain in #96
• feat: add KB mutation locks and atomic state writes by @gwokhou in #86
• fix: support ChatGPT-subscription & GitHub Copilot providers (litellm 1.87.2 + extra_headers) by @KylinMountain in #98
• fix(locks): cross-platform file locking — fix Windows fcntl crash by @KylinMountain in #99
• refactor(locks): delegate cross-platform file locking to portalocker by @KylinMountain in #100

New Contributors

• @moyu12-ae made their first contribution in #82
• @gwokhou made their first contribution in #85
• @cnndabbler made their first contribution in #90

Full Changelog: v0.3.0...v0.4.0

v0.4.0-rc3

What's Changed

• refactor(locks): delegate cross-platform file locking to portalocker by @KylinMountain in #100

Full Changelog: v0.4.0-rc2...v0.4.0-rc3

What's Changed

• refactor(locks): delegate cross-platform file locking to portalocker by @KylinMountain in #100

Full Changelog: v0.4.0-rc2...v0.4.0-rc3

v0.4.0-rc2

What's Changed

• fix(locks): cross-platform file locking — fix Windows fcntl crash by @KylinMountain in #99

Full Changelog: v0.4.0-rc1...v0.4.0-rc2

What's Changed

• fix(locks): cross-platform file locking — fix Windows fcntl crash by @KylinMountain in #99

Full Changelog: v0.4.0-rc1...v0.4.0-rc2

v0.4.0-rc1

What's Changed

Features

• feat(deck): HTML slide-deck generator + universal skill loader by @KylinMountain in #68
• feat: entity pages — first-class entities/type (configurable) + recompile backfill by @KylinMountain in #78
• feat: add KB mutation locks and atomic state writes by @gwokhou in #86
• fix: support ChatGPT-subscription & GitHub Copilot providers (litellm 1.87.2 + extra_headers) by @KylinMountain in #98

Bug Fixes

• fix: eliminate duplicate wiki-tool implementations in skill/tools.py by @wutongyuonce in #79
• fix: sync plugin manifest version to v0.3.0 by @moyu12-ae in #82
• fix: harden long PDF page extraction by @gwokhou in #85
• fix(compiler): don't cap concepts-plan at max_tokens=2048 (empty plans on reasoning models) by @cnndabbler in #90
• fix(cli): close litellm async clients after each compile/index (CLOSE-WAIT/FD leak) by @cnndabbler in #91
• refactor(cli,compiler): close litellm async clients inside the compile loop by @KylinMountain in #95
• fix: collision-resistant doc_name — same-stem documents no longer overwrite each other by @KylinMountain in #96

New Contributors

• @moyu12-ae made their first contribution in #82
• @gwokhou made their first contribution in #85
• @cnndabbler made their first contribution in #90

Full Changelog: v0.3.0...v0.4.0-rc1

Note: #95 is a refactor but addresses the same FD leak issue as #91, so I placed it under Bug Fixes for release clarity. Feel free to move it if you prefer.

Good luck with the release!