selfhostyourtech/reflector
2
0
Fork 0
mirror of https://github.com/Monadical-SAS/reflector.git synced 2026-04-24 06:05:19 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix: correct PRD goal (demo/eval, not dev replacement) and processor naming

Browse Source
This commit is contained in:
Igor Loskutov
2026-02-10 16:07:16 -05:00
parent 33a93db802
commit d0af8ffdb7
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
docs/01_ollama.prd.md
View File

@@ -4,7 +4,7 @@
Reflector currently uses a remote LLM endpoint (configurable via `LLM_URL`) for all post-transcription intelligence: topic detection, title generation, subject extraction, summarization, action item identification. The default model is `microsoft/phi-4`. Reflector currently uses a remote LLM endpoint (configurable via `LLM_URL`) for all post-transcription intelligence: topic detection, title generation, subject extraction, summarization, action item identification. The default model is `microsoft/phi-4`.
**Goal**: Run all LLM inference locally on developer machines (and optionally in self-hosted production), eliminating dependence on external LLM API providers. Zero cloud LLM costs, full data privacy, offline-capable development. One setup script, then `docker compose up` works. **Goal**: Let users run Reflector locally for demo/evaluation purposes without requiring cloud LLM API keys. A single setup script provisions a local LLM via Ollama, so the full post-transcription pipeline (topics, summaries, titles, etc.) works out of the box on a user's machine.
--- ---
@@ -28,7 +28,7 @@ All LLM calls go through one `LLM` class wrapping LlamaIndex's `OpenAILike` clie
### Call flow ### Call flow
``` ```
Hatchet workflows / Legacy processors Hatchet workflows / Celery processors
-> LLM.get_response() or LLM.get_structured_response() -> LLM.get_response() or LLM.get_structured_response()
-> LlamaIndex TreeSummarize + StructuredOutputWorkflow -> LlamaIndex TreeSummarize + StructuredOutputWorkflow
-> OpenAILike client (is_chat_model=True, is_function_calling_model=False) -> OpenAILike client (is_chat_model=True, is_function_calling_model=False)