for CTOs · technical founders · engineering leaders
Find what slipped between the conversation and the merge.
Threadout reconciles one recent initiative across your communication exports, project tracker, and GitHub. You receive a human-reviewed report showing commitments that were never tracked, became buried, or were delivered differently than agreed.
one initiative · bounded read-only inputs · evidence linked to every finding · deletion receipt after delivery
the signature catch
“Cover up user names in replays” — reverted the same day it was asked.
A founder asks for it in a walkthrough. A code change that afternoon quietly undoes the masking. No ticket ever exists. The reconciliation catches it from the transcript, the two diffs, and a replay sample — the ask only lives in a conversation, so no code-and-analytics tool can see this gap.
- disposition
- human review required — the finding stays held until a human confirms the mismatch
why we built this — founder-problem proof
We built Threadout after manually reconstructing the chain between 34 meeting transcripts, 1,016 tracked tasks, and the implementation evidence behind them.
the recurring failures: scope drift · unsupported remediation claims · lost deliverables · permission blockers
not customer proof — our own. the scan productizes that reconstruction.
illustrative integrity-scan scenario · not a customer outcome · the ask only exists in a transcript
what threadout catches
dropped
Agreed out loud, never tracked. The ask lives only in a conversation — three weeks later someone asks whatever happened to it.
buried
Captured, then displaced. The ticket is technically alive and effectively invisible under newer priorities.
intentionally parked
Deliberately deferred — but the reason was never recorded, so it keeps resurfacing and getting relitigated.
delivered differently
It shipped — but not as decided. Nobody re-checks the meeting against the merge. The scan does.
// AI made shipping faster. It didn't make anything easier to trace — it made more of it.
what you receive, and how the scan produces it
what you receive
- Executive summary — the three-to-five findings that matter, ranked by consequence.
- Commitment ledger — every ask in the window: said → captured → shipped → verified.
- Findings with source links — each one traceable to the transcript line, ticket, or diff.
- Recommended actions — exact change, scope, and acceptance criteria per finding.
- Coverage receipt — exactly what was read, what wasn't, and what that means.
- Deletion receipt — proof your data left when the engagement closed.
how the scan works
- Define one initiative and the data boundary together.
- You provide exports or bounded read-only evidence — communication export, one tracker project, up to three repos.
- Threadout reconciles communication, planning, and implementation in an isolated environment.
- Every finding is human-reviewed before you see it.
- You receive the report and an executive readout.
- Data is deleted per the agreed policy — receipt included.
connected design partners receive the same reconciliation continuously — by invitation, after the scan.
how the connected workspace works, once you are in it
01 · connect
Read-only connectors, your stack
Slack public and invited channels only — never DMs — your tracker (Jira, Linear, or Asana), and GitHub. Connectors are being built read-only and scoped, with a sources page that will show exactly what the engine can see.
planned connectors: slack · jira · linear · asana · github — notetakers next: otter · fathom · granola
02 · reconcile
The engine builds the full picture
Every ask becomes a thread: what was said, what was decided, what got ticketed, what merged, what production does now. Gaps between stages are computed — with evidence, never vibes.
said → captured → shipped → verified · every claim shows its source
03 · digest
One weekly email that calls it out
What was dropped, what's buried, what was chosen not to do, what shipped as promised. Ranked by risk, one click to the proof. That's the whole beta — and it's enough.
what lands in your inbox
two ways to start
Managed Integrity Scan
One initiative, fully reconciled. You hand us exports or bounded read-only access; we process in an isolated environment, human-review every finding, and deliver the evidence-backed report. Fixed fee. Deletion receipt when we're done.
slack export · tracker export · github read-only · nothing touches a shared runtime
book a scan ▸Connected workspace
What the private beta wires up: your sources, read-only, in an isolated workspace of your own — no shared runtime, no shared database — continuous reconciliation, and a Monday digest that arrives on its own. Not open yet: one design partner onboarded at a time, from the waitlist.
planned: read-only connectors · weekly digest · land / kill / assign / dispute
join the waitlist ▸trust & data handling
- What data does a scan need?
- A bounded set you choose: one communication export (channels you pick — never DMs unless you include them), one tracker project, up to three repositories. Nothing else.
- Is access read-only?
- Yes — exports or read-only tokens only. Scans produce a report; they never write to your systems, create tickets, or take automatic actions.
- Does Threadout train on our data?
- No. Customer evidence is used only to perform the contracted scan. Model-provider handling and retention terms are disclosed in the engagement agreement.
- Who sees the findings?
- A human reviews every finding before delivery, and findings describe work and systems — never person-level scores. You control distribution.
- How long is data retained?
- For the engagement, then deleted per the agreed policy. You get a deletion receipt covering evidence, working files, and report inputs.
- What's live today?
- The managed scan: isolated processing, bounded inputs, human review, deletion receipt. The connected workspace is in a stage-gated private build — its tenant-isolation controls are checked before any customer data connects.
trust claims above are specific to the managed scan — we don't claim connected-beta controls that aren't deployed yet.
start here — the one thing you can buy today
founding engagement · fixed fee · manually invoiced
- · one recent initiative
- · one bounded communication source set
- · one tracker project
- · up to three GitHub repositories
- · evidence-backed finding report
- · executive readout
- · deletion receipt
A recurring, isolated, read-only workspace with the human-reviewed weekly digest. Scoped and priced separately — it carries ongoing infrastructure, connector operations, and security work. Scan customers get first invitations.
later: the continuous threadout workspace — the self-service product, once the pilot earns it.
the connected workspace, when it opens — proposed pricing
The scan above is what you can buy today. This is where the subscription lands once the pilot earns it: flat per workspace, tiered by connected sources. Value accrues to the org, not seats — so we don't charge for seats. Everyone on the team gets in.
proposed pricing — no checkout or paid subscription is available yet · waitlist only
design partner
$0 during beta
3 connectors, wired with you · weekly digest · your feedback shapes the engine
join the waitlist ▸workspace
$359 / month
up to 8 connectors · unlimited seats · weekly digest + attention + threads · read-only MCP for coding agents — later
start — request access ▸scale
custom
unlimited sources · SSO · retention controls · audit exports · counterparty review links
request a demo ▸no per-seat pricing · proposed terms — nothing is purchasable yet