Loose Ends: a local AI that rereads my notes so I do not have to

Description and use

Loose Ends scans local notes, extracts tasks, and produces a checklist using a local LLM, keeping data on the machine.

Problem

Maintaining a separate task list requires rereading notes, a low‑energy chore that is often postponed.

Outcome

The tool turns two notes into three actionable checklist items, merges a duplicate promise, and ignores a completed task. A second run reuses the cached report with no model inference.

Workflow

• Scan a folder of .md and .txt files
• Label each line with filename:line
• Send labeled text to a local Ollama model (qwen2.5-coder:7b-instruct-q4_K_M) with a prompt for structured JSON
• Validate that each suggestion has a real source token
• Group items by today, this_week, later, or unknown
• Render the checklist to a markdown file
• Cache the report using SHA‑256; reuse if notes unchanged

Evidence

Loose Ends scans a folder of Markdown or text notes and produces one reviewable checklist.

Source

The slow-and-steady part is the rerun. When the notes have not changed, a SHA-256 cache returns the last report with zero model calls.

Source

The Python CLI uses only the standard library and a local Ollama endpoint. No cloud key is needed.

Source