Meetings captured silently. Documents woven into a searchable knowledge base. Three knowledge tiers, ranked by quality. Your AI briefed instantly. Your unfair advantage.
Public sign-up is not open yet.
Meetings dissolve into air. Decisions scatter. Most teams forget 80% of conversational detail within 48 hours. Your colleagues who seem to remember everything aren't smarter—they just capture it. That's the gap between keeping up and staying ahead.
"We're moving to a multi-phase rollout in Q3, and I think we need to re-evaluate the pricing model before we—"
Lost by Wednesday. Connected across three weeks by Haberdash.
For people who need institutional memory without taking notes.
The moment a meeting ends, Haberdash has already surfaced what matters: key points, decisions, action items—each confidence-scored. Who said what. What needs to happen next. No summaries, no fluff. Just extracted fact.
Product Strategy Sync
Team aligned on Q3 pricing restructure. Julian flagged hemline tension in new prototype — double-stitch reinforcement proposed, awaiting textile stress tests.
Decisions
Actions
Week 42 — October Cycle
“Pricing Model Revision (continuing from W41): consensus reached on 15% margin — resolving the open question from three weeks ago.”
Prior Week Follow-up
Every week, a synthesis pipeline connects your meetings into cross-cutting themes, tracks how priorities shift week-to-week, and surfaces narrative arcs. A decision made in Week 12 surfaces when it matters again in Week 15. Automatically.
Upload documents, training materials, research, policy—whatever you need to remember. Haberdash discovers topics across sources and weaves comprehensive 30,000-word references. As new knowledge arrives, themes update automatically. Your institutional memory, living.
Sections
Updated 2 days ago — 3 new sources incorporated
Recent Development
W42 sync confirmed 15% margin consensus. Section 4 updated automatically.
Grant Claude, ChatGPT, or any AI client direct access to your knowledge base via MCP tools. Quality-tiered search: themes first, then meetings, then reference notes. Your AI stays informed as context evolves, picking up exactly where you left off. No manual context pasting.
> haberdash_briefing(topic: "atelier pricing")
Briefing exists (updated W41). 3 new sources since last update.
Update: Team reached consensus on 15% margin in W42 sync. Julian's stress tests confirmed reinforcement viability. One open item remains: quarterly indexing formula.
_
Haberdash operates invisibly. Meetings capture and transcribe themselves. Documents you upload are indexed automatically. Everything flows into three knowledge tiers — meetings, themes, and reference materials. Your AI searches by quality: themes first, then meetings, then short-term notes. No more fumbling for context.
Three tiers of knowledge, ranked by relevance. Themes first—your institutional memory. Then meetings. Then reference notes. Search always returns what matters most.
Comprehensive living documents woven from your uploaded materials and meeting insights. 30,000–50,000 words per theme. Updated automatically as new knowledge arrives. Your institutional memory—what your AI cites first.
Captured, transcribed, and enriched. Every decision and action item confidence-scored. Weekly synthesis connects patterns across meetings and surfaces narrative arcs—what shifted from last week, what's unresolved.
Training docs, handover guides, policies, reference materials. You control what gets indexed. Time-decayed relevance keeps search focused. Lower-ranked than themes, but searchable when needed.
From raw audio to confidence-scored decisions in under 30 seconds. Cross-week intelligence in under 5 minutes.
Meeting audio
with notice
where required.
Whisper V3.
Speakers
identified.
Decisions.
Actions.
Confidence-scored.
Themes.
Narrative arcs.
Across weeks.
Fully local. Your keys. Your machine.
Cloud scale. No API keys.
Enterprise scale. Total control.