OKRs your team actually keeps up with — with an AI coach in the loop.
A weekly check-in habit, an AI coach that reads every plan against your goals, and managers who finally see the whole picture. Self-hosted, open-source, your data.
What it does
Quarterly OKRs, top to bottom
Set goals from CEO to IC with hierarchical cascade and cross-functional alignment as first-class citizens — not an afterthought.
~90-second weekly check-ins
Every Monday each person submits a 3–7 item plan. Low friction by design, so the habit actually sticks.
An AI coach on every plan
The LLM reads each plan against your KRs, recent history, and your org’s Soul.md, then gives structured feedback on alignment, realism, and coverage.
Soul.md — your org’s voice
One per-org markdown file defines the coach’s tone, vocabulary, and what “good” looks like. The 4000-token cap is the feature, not a limit.
Meet people where they are
Notifications over Email, Slack, Telegram, and WhatsApp with an automatic fallback chain and per-user preference order.
Bilingual EN / ID
Every user picks a language. UI, notifications, and AI feedback all respect it — no second-class translation.
What it deliberately is not
- Not a Lattice replacement — no 1:1s, performance reviews, engagement surveys, or comp.
- Not a project manager — KRs are outcome metrics, not task lists.
- Not a data grab — self-hosted and single-tenant; your goals never leave your instance.
Own your goal data
AGPL-3.0 and self-hosted. Run it on your own infra, bring your own LLM key (Anthropic, OpenAI, or self-hosted Ollama), and keep every objective inside your walls. No telemetry, no shared state, no vendor lock-in.