Research Synthesis

Merge notes, papers, and meeting minutes into a single searchable corpus, then use AI to surface key themes.

The Problem

Research projects accumulate notes, papers, meeting minutes, and correspondence across months or years. Key insights get buried in the volume. You need to synthesise findings across all sources to identify patterns and themes.

Workflow

1. Configure All Research Sources

ingestion:
  sources:
    - type: diary
      path: "~/research/notes"
      entry_separator: "## "
    - type: meeting_note
      path: "~/research/supervisor-meetings"
    - type: document
      path: "~/research/papers"
    - type: document
      path: "~/research/reports"

2. Ingest and Explore

uv run python ingest.py --reset
 
# Understand what you have
uv run python explore.py
uv run python explore.py --source-type diary
uv run python explore.py --source-type document

3. Theme-Based Searches

Search for recurring themes across all source types:

mkdir -p evidence
 
uv run python query.py \
  --semantic "methodology approach research design" \
  --top-k 200 --export-json evidence/methodology.json
 
uv run python query.py \
  --semantic "findings results conclusions outcomes" \
  --top-k 200 --export-json evidence/findings.json
 
uv run python query.py \
  --semantic "limitations gaps future work recommendations" \
  --top-k 200 --export-json evidence/gaps.json
 
uv run python query.py \
  --semantic "key insight breakthrough discovery pattern" \
  --top-k 200 --export-json evidence/insights.json

4. Merge and Analyse

uv run python merge.py evidence/*.json --output evidence/merged.json
 
uv run python analyze.py evidence/merged.json \
  --full-pipeline \
  --context "Synthesise key themes, recurring findings, methodological patterns, and identify gaps or contradictions across the research corpus" \
  --model deepseek

5. Export for Writing

# Copy to clipboard for pasting into a writing tool
uv run python export.py analysis_output.md --clipboard
 
# Or save to file
uv run python export.py analysis_output.md --output synthesis-report.md

Tips

Use --year to compare themes across different time periods
Diary entries often capture informal insights that formal papers miss
Meeting notes contain decisions and direction changes not recorded elsewhere
For fully private analysis (no cloud), use --local with Ollama
Run multiple rounds with different context prompts to explore different angles

Contract Review Overview