A research thread manager for AI agents. Track topics over time, accumulate notes and sources, and pair with scheduled jobs to produce regular research digests.
Your agent picks the threads. Your agent does the searching. This tool keeps the state.
Ships: A zero-dependency Python CLI (scripts/deep-current.py) that manages research threads as local JSON.
Doesn't ship: Web search, link following, or report generation. Those come from your agent's own tools.
The CLI handles what to research. Your agent handles how.
# Create a thread
python3 scripts/deep-current.py add "Carnivore Diet Research"
# Add notes, sources, findings as you go
python3 scripts/deep-current.py note carnivore "New study on protein satiety in women"
python3 scripts/deep-current.py source carnivore "https://example.com/study" "2024 protein satiety meta-analysis"
python3 scripts/deep-current.py finding carnivore "High-protein diets show 25% better satiety scores"
# See what you're tracking
python3 scripts/deep-current.py list
python3 scripts/deep-current.py show carnivore
python3 scripts/deep-current.py digest| Command | Purpose |
|---|---|
list |
Show all threads with status |
show <id> |
Full thread details |
add <title> |
Create new thread |
note <id> <text> |
Add dated research note |
source <id> <url> [desc] |
Add source/reference |
finding <id> <text> |
Record key finding |
status <id> <active|paused|resolved> |
Change thread status |
digest |
Summary of all active threads |
decay |
Prune stale threads (>90 days inactive) |
Thread IDs are auto-generated slugs. Prefix matching works (carn matches carnivore-diet-research).
Install from ClawHub:
clawhub install deep-currentSchedule a nightly cron job that tells your agent to pick threads, research them with web_search/web_fetch, and write findings to deep-current-reports/YYYY-MM-DD.md. See SKILL.md for the full cron prompt template.
The CLI is framework-agnostic. Any agent that can:
- Run shell commands (to call the CLI)
- Search the web (to do the actual research)
- Write files (to output reports)
...can use Deep Current. Point your agent at the CLI, give it a prompt like "pick a thread, research it, write findings," and you're set.
It's just Python with no dependencies. Use it as a personal research tracker without any agent at all:
python3 scripts/deep-current.py add "Topic I'm curious about"
python3 scripts/deep-current.py note topic "Found an interesting paper on..."
python3 scripts/deep-current.py digestEverything lives in deep-current/currents.json — a single JSON file. Back it up, version it, move it between machines. Reports go to deep-current-reports/ as individual dated markdown files.
MIT