v0.4 · 325 tests passing

An AI co-pilot for boring ops work in retail teams.

Ingest orders & customer data, score churn and demand, recommend the next action, and dispatch to Slack, email, or Sheets — every plan graded against a 55-case eval corpus.

live demo · ~5s response · Heuristic + Gemini backends · CI-pinned baseline
$ liveops ask
› which 5 products bring in the most revenue?

plan:  top_products(n=5)
trace: top_products → 5 rows in 0.04s
answer:
  · SKU-247  $182,431
  · SKU-019  $151,902
  · SKU-118  $138,544
  · SKU-882  $129,107
  · SKU-014  $118,392

› which customers are about to churn?

plan:  churn_risk(days=60)
trace: 1,204 customers · 18% At-Risk · 7% Churned
325
tests passing
10
analyst tools
55
eval cases
4
data connectors
capabilities

Detect, recommend, and ship — eval-graded.

Two product surfaces share a single tool registry. Both code paths are scored on the same 55-case corpus, so improvements show up everywhere a tool is used.

Ops agent

Detect → Recommend → Act

Watches your KPIs, flags incidents with anomaly + change-point detection, and ships the right next step to Slack, email, or Sheets. A Thompson bandit learns from your feedback.

  • · Anomaly + drift detection per metric
  • · Bandit-driven recommender across 5 action templates
  • · Slack / email / Google-Sheets dispatch
  • · Bcrypt + Fernet encrypted credentials
Analyst workbench

Ten tools, eval-graded

Plain-English questions over orders & customers. Plans are scored on tool-match, args-match, forbidden-tool avoidance, and must-succeed execution.

  • · Revenue trends · top products · top customers
  • · RFM segments · BCG quadrants · co-purchases
  • · Price elasticity · churn risk · cohort retention
  • · Heuristic + Gemini backends, both eval-tested
pinned baselines · honest signal

Two corpora, three numbers.

The headline 100% is the heuristic against questions it was tuned on — useful for catching regressions, not for proving generalization. The held-out split paraphrases each question with vocabulary the heuristic has never seen, so we can see how each backend behaves on novel input. Both numbers are real and pinned in tests/fixtures/.

View harness
main · heuristic tuned
100.0%
55 / 55 cases

Regression-catcher. Same regex, same questions — moves only when something breaks.

holdout · heuristic novel
7.1%
1 / 14 paraphrases

"Identify my whales", "bucket my buyers" — none of the regex's keywords appear. The heuristic falls off a cliff.

holdout · LLM generalization
28.6%
4 / 14 paraphrases

Same paraphrases through Gemini. ~4× the heuristic — meaningful generalization, with room to grow.

end-to-end pipeline

Nine stages. All observable.

Every stage is restartable, instrumented, and eval-checked.

01 Ingest
CSV · Parquet · Postgres · S3 · Sheets
02 Clean
Schema infer · null fill · type coerce
03 Score
Anomaly · drift · churn · demand
04 Plan
Heuristic / Gemini tool router
05 Decide
Bandit picks next action
06 Notify
Slack · email · Sheets
07 Audit
Plan + obs trace persisted
08 Eval
55 cases on every PR
09 Deploy
Docker · Render · GHCR