Commit Graph

Select branches

Hide Pull Requests

chore/finish-llm-provider-abstraction-epic

feat/ltx-workflow-canvas-format

gh-pages

main

v0.1.0

Mono Color

• ccb1cb0bcc Deployed f16da57 to 0.1.1.dev88+gf16da579b with MkDocs 1.6.1 and mike 2.1.3 gh-pages github-actions[bot] 2026-07-29 22:44:02 +00:00
• f16da579bc feat(ui): add output-text previews and live retry status to comfydv nodes main James VeitchandClaude Sonnet 5 2026-07-29 23:43:13 +01:00
• 4630df6702 Deployed 7a40d22 to 0.1.1.dev87+g7a40d2223 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-29 21:20:13 +00:00
• 7a40d22230 feat(llm): escalate retry timeouts, expose seed_used, add retry UI feedback James VeitchandClaude Sonnet 5 2026-07-29 22:19:22 +01:00
• 0b81194967 Deployed 42c614d to 0.1.1.dev86+g42c614d5f with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-29 16:45:51 +00:00
• 42c614d5f8 feat(llm): add user-configurable custom_phrases to refusal detection James VeitchandClaude Sonnet 5 2026-07-29 17:44:32 +01:00
• 22b0ba5f4f Deployed a0f391f to 0.1.1.dev85+ga0f391f21 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-29 16:35:40 +00:00
• a0f391f212 fix(llm): catch bare "cannot generate" phrasing without a leading "I" James VeitchandClaude Sonnet 5 2026-07-29 17:34:41 +01:00
• f0889b8d9c Deployed 3c5136a to 0.1.1.dev84+g3c5136af7 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-29 16:31:13 +00:00
• 3c5136af70 fix(llm): stop gating the embedding refusal check on is_ambiguous James VeitchandClaude Sonnet 5 2026-07-29 17:30:19 +01:00
• c48d12d405 Deployed 97249ee to 0.1.1.dev83+g97249ee28 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-29 16:17:39 +00:00
• 97249ee281 fix(llm): catch "I am unable to" phrasing in refusal detection James VeitchandClaude Sonnet 5 2026-07-29 17:16:38 +01:00
• c1be110d29 Deployed 1f138e5 to 0.1.1.dev82+g1f138e599 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-29 16:11:58 +00:00
• 1f138e5991 feat(llm): detect refusal/deflection responses and retry with a new seed (#36) James VeitchandClaude Sonnet 5 2026-07-29 17:11:08 +01:00
• f8114f01f4 feat(workflows): add fully-wired canvas-format LTX pipeline feat/ltx-workflow-canvas-format James VeitchandClaude Sonnet 5 2026-07-26 09:07:31 +01:00
• 90b90f4648 Deployed 2724f53 to 0.1.1.dev81+g2724f535f with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-26 07:16:40 +00:00
• 2724f535f5 fix(workflows): disable thinking by default in LTX-2.3 pipeline (#34) James VeitchandClaude Sonnet 5 2026-07-26 08:15:49 +01:00
• d29d399100 Deployed a613006 to 0.1.1.dev80+ga613006f5 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-25 22:45:54 +00:00
• a613006f5a fix(ollama): bypass stale combo validation for LLM model widgets (#33) James VeitchandClaude Sonnet 5 2026-07-25 23:45:02 +01:00
• 61753a4fb8 Deployed 56c7d4c to 0.1.1.dev79+g56c7d4ca5 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-25 09:39:09 +00:00
• 56c7d4ca56 Merge pull request #32 from darth-veitcher/feat/disable-thinking-toggle James Veitch 2026-07-25 10:38:17 +01:00
• cf7cfeb6aa feat(ollama): add disable-thinking toggle for both providers James VeitchandClaude Sonnet 5 2026-07-25 10:37:46 +01:00
• 7e83e43da8 Deployed c9aefef to 0.1.1.dev77+gc9aefef91 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-25 07:26:55 +00:00
• c9aefef915 Merge pull request #31 from darth-veitcher/fix/ollama-structured-output-native-format James Veitch 2026-07-25 08:26:04 +01:00
• f060698e7f feat(workflows): add LTX-2.3 multi-agent I2V pipeline example James VeitchandClaude Sonnet 5 2026-07-25 08:25:12 +01:00
• 0f4b2234db fix(ollama): route structured output through native /api/chat, fix nullable field validation James VeitchandClaude Sonnet 5 2026-07-25 08:25:01 +01:00
• c71a57c218 Deployed ecba529 to 0.1.1.dev74+gecba52999 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-24 21:41:21 +00:00
• ecba52999d Merge pull request #30 from darth-veitcher/feat/jinja2-fromjson-filter James Veitch 2026-07-24 22:40:32 +01:00
• 24119e3ab6 feat(format-string): add fromjson Jinja2 filter James VeitchandClaude Sonnet 5 2026-07-24 22:40:08 +01:00
• 305638be3d Deployed 1f30c2f to 0.1.1.dev72+g1f30c2ff2 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-24 21:01:08 +00:00
• 1f30c2ff2f Merge pull request #29 from darth-veitcher/fix/jinja2-template-variable-extraction James Veitch 2026-07-24 22:00:15 +01:00
• ecfc7eca59 fix(tests): pin comfydv module resolution before pytest fixture setup James VeitchandClaude Sonnet 5 2026-07-24 21:59:38 +01:00
• e534fac121 fix(format-string): extract Jinja2 template variables via AST, not regex James VeitchandClaude Sonnet 5 2026-07-24 21:59:23 +01:00
• e6bd58815a Deployed 6c3c341 to 0.1.1.dev69+g6c3c34191 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-23 11:33:23 +00:00
• 6c3c341913 Merge pull request #24 from darth-veitcher/claude/chatcompletion-image-input-7mm6q5 James Veitch 2026-07-23 12:32:31 +01:00
• 39887e6b5b Deployed 8cb9a74 to 0.1.1.dev49+g8cb9a7446 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-23 07:00:16 +00:00
• 8cb9a74462 Merge pull request #28 from darth-veitcher/fix/ollama-incomplete-response-race James Veitch 2026-07-23 07:59:27 +01:00
• 5e678c2700 fix(ollama): raise a clear error on Ollama's incomplete done:false stub James VeitchandClaude Sonnet 5 2026-07-23 07:58:31 +01:00
• 3854b27cba fix(009): surface a clear error for non-vision Ollama models given an image James VeitchandClaude Sonnet 5 2026-07-23 07:42:57 +01:00
• 54467a78cb Deployed e432054 to 0.1.1.dev47+ge432054b4 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-22 20:47:16 +00:00
• e432054b47 Merge pull request #26 from darth-veitcher/fix/dockerfile-test-python-313 James Veitch 2026-07-22 21:46:27 +01:00
• 09fe013ff9 fix(tests): update Dockerfile test to expect python:3.13-slim James VeitchandClaude Sonnet 5 2026-07-22 20:20:33 +01:00
• 0f73f3c2cc chore(009): register build bullet, record quality-gate status (T009) Claude 2026-07-22 19:11:50 +00:00
• 24917be802 docs(009): document image input on Chat Completion (T008) Claude 2026-07-22 19:10:10 +00:00
• 46c14a872d feat(009): structured-output image input via BinaryContent (T007-I) Claude 2026-07-22 19:08:19 +00:00
• 0cfe0d22ff test(009): failing tests for structured-output image input (T007-T) Claude 2026-07-22 19:07:39 +00:00
• bc7b33922b feat(009): map images to OpenAI content-parts for llama.cpp (T006-I) Claude 2026-07-22 19:06:40 +00:00
• 1028315d37 test(009): failing tests for llama.cpp image content-parts (T006-T) Claude 2026-07-22 19:05:56 +00:00
• a2f2464fdb feat(009): forward images via Ollama /api/chat, byte-identical text path (T005-I) Claude 2026-07-22 19:05:20 +00:00
• 4a631cc8dc test(009): failing test for Ollama image passthrough (T005-T) Claude 2026-07-22 19:05:00 +00:00
• fd9054c03f feat(009): optional IMAGE input on ChatCompletion (T004-I) Claude 2026-07-22 19:04:06 +00:00
• 23d96f1559 test(009): failing tests for ChatCompletion IMAGE input (T004-T) Claude 2026-07-22 19:03:24 +00:00
• 9ece5c0efa feat(009): _encode_image_tensor for ChatCompletion image input (T003-I) Claude 2026-07-22 19:02:36 +00:00
• e225dcaaa3 test(009): failing tests for _encode_image_tensor (T003-T) Claude 2026-07-22 19:02:06 +00:00
• 1c72e0b2ea feat(009): add optional images carrier to Message (T002-I) Claude 2026-07-22 19:01:00 +00:00
• 7e56f2e5bf test(009): failing tests for Message.images carrier (T002-T) Claude 2026-07-22 19:00:21 +00:00
• f4a3390f13 build(deps): add pillow to dev group for image-encode tests (T001) Claude 2026-07-22 18:59:57 +00:00
• a40c1ba373 docs(tasks): task breakdown + BDD witnesses for VLM image input (009) Claude 2026-07-22 18:56:26 +00:00
• eab86b0324 docs(plan): plan VLM image input (009) — research, data-model, contracts Claude 2026-07-22 18:50:09 +00:00
• 45e49a98eb docs(spec): draft VLM image input for ChatCompletion (009) Claude 2026-07-22 18:38:39 +00:00
• eb236eafdd Deployed 87ec79b to 0.1.1.dev45+g87ec79be1 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-20 08:04:03 +00:00
• 87ec79be1b Merge pull request #23 from darth-veitcher/fix/retry-blank-output-with-seed-and-backoff James Veitch 2026-07-20 09:03:16 +01:00
• 1dcef77a2b fix: keep chat_structured retry's nested seed in sync with the top-level one James VeitchandClaude Sonnet 5 2026-07-20 08:59:06 +01:00
• 09a173e552 fix: retry blank/failed LLM responses with a new seed and backoff James VeitchandClaude Sonnet 5 2026-07-20 08:49:35 +01:00
• ad32d98a5e Deployed 92da4fc to 0.1.1.dev42+g92da4fc11 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-12 09:34:37 +00:00
• 92da4fc117 Merge pull request #22 from darth-veitcher/fix/run-async-real-comfyui-event-loop James Veitch 2026-07-12 10:33:50 +01:00
• 1fe3bef663 docs(tests): stop overclaiming what the _run_async regression tests prove James VeitchandClaude Sonnet 5 2026-07-12 09:39:55 +01:00
• 3bd857f0ab fix: chat_structured() crashes with real asyncio.run() error in live ComfyUI James VeitchandClaude Sonnet 5 2026-07-12 09:36:21 +01:00
• ca09b99e5b Deployed 834ae3e to 0.1.1.dev39+g834ae3e9e with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-11 20:56:00 +00:00
• 834ae3e9e3 Merge pull request #18 from darth-veitcher/chore/finish-llm-provider-abstraction-epic James Veitch 2026-07-11 21:55:08 +01:00
• 2df9e87838 chore: resolve merge conflicts with main chore/finish-llm-provider-abstraction-epic copilot-swe-agent[bot] 2026-07-11 20:47:06 +00:00
• 0012840fc6 Deployed 27b26ec to 0.1.1.dev35+g27b26ec7a with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-11 17:11:27 +00:00
• 27b26ec7a4 Merge pull request #21 from darth-veitcher/fix/comfyui-relative-import-loading James Veitch 2026-07-11 18:09:56 +01:00
• 16b7e944f0 fix(tests): make the import-compat proof order-independent James VeitchandClaude Sonnet 5 2026-07-11 18:02:57 +01:00
• 36e2cd123c docs: rebuild README around live screenshots, fix stale doc tooling James VeitchandClaude Sonnet 5 2026-07-11 17:48:37 +01:00
• 1e9a8a3b6a fix: comfydv failed to load in real ComfyUI since PR #17 (critical) James VeitchandClaude Sonnet 5 2026-07-11 17:48:20 +01:00
• 6c3c0fc485 Deployed d59a85d to 0.1.1.dev31+gd59a85dba with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-11 15:48:58 +00:00
• d59a85dba6 Merge pull request #19 from darth-veitcher/008-llamacpp-integration James Veitch 2026-07-11 16:48:09 +01:00
• 15bb25efa8 test(llamacpp): live-verify chat_structured(), strengthen idempotency tests James VeitchandClaude Sonnet 5 2026-07-11 16:24:52 +01:00
• 282198a7df fix(llamacpp): make load_model/unload_model actually idempotent James VeitchandClaude Sonnet 5 2026-07-11 16:10:51 +01:00
• ac248e1e99 chore(beacon): regenerate ROADMAP.md rollup James VeitchandClaude Sonnet 5 2026-07-11 15:49:27 +01:00
• ea66ceee9d docs(beacon): flip llamacpp-integration epic Planning->Active James VeitchandClaude Sonnet 5 2026-07-11 15:48:03 +01:00
• cf3f7fc174 fix(llamacpp): surface a clear error for non-router-mode servers (FR-006) James VeitchandClaude Sonnet 5 2026-07-11 15:43:12 +01:00
• 5af502d07e feat(llamacpp): add LlamaCppProvider + LlamaCppClient node (spec 008) James VeitchandClaude Sonnet 5 2026-07-11 15:27:29 +01:00
• d3ee35bfda docs(beacon): DESIGN artifacts for llama.cpp Model Integration (spec 008) James VeitchandClaude Sonnet 5 2026-07-11 15:17:54 +01:00
• 27916f4fb0 chore(beacon): regenerate ROADMAP.md gantt/table for llamacpp-integration Active James VeitchandClaude Sonnet 5 2026-07-11 15:11:25 +01:00
• 689aa2a391 chore(beacon): finish llm-provider-abstraction epic James VeitchandClaude Sonnet 5 2026-07-11 15:10:57 +01:00
• 736374296d Deployed 9ca65bc to 0.1.1.dev23+g9ca65bc29 with MkDocs 1.6.1 and mike 2.1.3 github-actions[bot] 2026-07-11 14:10:04 +00:00
• 9ca65bc29a Merge pull request #17 from darth-veitcher/007-llm-provider-abstraction James Veitch 2026-07-11 15:08:14 +01:00
• c32b6a7e42 docs(beacon): note PR #17 open on the llm-provider-abstraction epic James VeitchandClaude Sonnet 5 2026-07-11 14:42:59 +01:00
• 3df38485d2 docs: update README/docs for the generic node rename, add migration table (FR-009) James VeitchandClaude Sonnet 5 2026-07-11 14:42:19 +01:00
• 8e227790a9 fix(llm): forward options through to structured output (beacon-reviewer finding) James VeitchandClaude Sonnet 5 2026-07-11 14:35:51 +01:00
• 6c2c906c4a docs(beacon): mark tasks.md Phase 7 subsumed by T-CUT-10/12 James VeitchandClaude Sonnet 5 2026-07-11 14:29:20 +01:00
• 1879ca2265 test(llm): complete atomic node cutover — test_ollama.py rewrite (T-CUT-08/10/12) James VeitchandClaude Sonnet 5 2026-07-11 14:28:34 +01:00
• d90ae98fed feat(llm): atomic node cutover — generic ChatCompletion/LLM* nodes (T-CUT-01/04-07/11) James VeitchandClaude Sonnet 5 2026-07-11 14:19:07 +01:00
• fb111c0142 feat(llm): implement OllamaProvider method bodies (T-CUT-02/03) James VeitchandClaude Sonnet 5 2026-07-11 14:14:51 +01:00
• 22dbddc8eb docs(beacon): fix tasks.md checkbox hygiene, add trackable T-CUT phase James VeitchandClaude Sonnet 5 2026-07-11 14:12:02 +01:00
• ef2464aebb docs(beacon): properly spec the atomic node cutover (issue #16) James VeitchandClaude Sonnet 5 2026-07-11 13:55:39 +01:00
• 736ffd1e77 feat(llm): shared chat_structured() via pydantic-ai (T011-T013) James VeitchandClaude Sonnet 5 2026-07-11 13:37:16 +01:00
• 1b1c7830d5 docs(beacon): correct tasks.md — US1/US3 cutover is atomic, not incremental James VeitchandClaude Sonnet 5 2026-07-11 13:31:21 +01:00