Workflows

Research Workflows

Research Workflow

Run Search, DailyPaper, and Analyze.

3 topicspapers.cool · arXiv API · HF DailyarXiv · VenueLLM onJudge on

Current phase

Idle

Not run yet

Candidates

0

No data yet

Research handoff

Global

0 queued · 0 alerts

01

Search & collect

Pending

3 topics · 3 sources

02

Build DailyPaper

Pending

Top N 10 · files saved

03

Analyze & hand off

Pending

0 judged · 0 delivery channels

Run setup

Run setup

Topics

1
2
3

Sources

Branches

LLM analysis
Judge scoring
Save artifacts
Email digest
Newsletter push

Ready to run.

Run Snapshot

Current workflow state

Idle

Last updated

Not run yet

Output mode

Persistent artifacts

Judge coverage

0 / 0 papers

Delivery

Manual review only

Stage progress0%

Run Search.

Delivery & Handoff

Keep the chain connected

Output: ./reports/dailypaperDelivery disabled

Open Research Workspace

Hand off reviewed papers to Research for deeper exploration and feed curation.

Browse Knowledge Base

Use Wiki as the grounding layer for metrics, architectures, and methods before delivery.

Workflow Playbook

Attach docs and wiki to the operator loop

A reasonable chain is: topic selection inside Dashboard, concept grounding in Wiki, evidence review in Research, and then delivery from the same run snapshot.

Judge Metrics in Wiki

Open the concept browser filtered to evaluation and judge-related entries.

Architecture Concepts in Wiki

Use the knowledge base to refresh related methods before deciding what to save or push.

Open Papers Library

Move from the dashboard run into saved papers and reference exports once the shortlist is stable.

Tune Providers & Delivery

Route models and configure push channels before turning this workflow into a production ritual.

3
Queries
0
Papers Found
0
Judged
Idle
Phase
Workflow DAG

0 papers

Run Search to find papers, then DailyPaper to rank and compose a report, then Analyze to run Judge/Trends.