Menu
ProductReplay & DevToolsSignalsCall qualityWhy (AI)Live AssistCanvas recordingPrivacy & complianceSelf-hosted
SolutionsHealthcare & virtual careEngineeringSupportProduct teamsKiosks & monitoring
MorePricingDocsCompareTrust centerChangelogGlossaryAboutContactPress kit
Start free Talk to sales Log in

Root causes, with receipts.

Every claim cites a moment in the replay, and the Player turns that citation into a jump link.

A claim without a citation is never rendered.

Signals 0 fired

    Why generated from this session

    Loading explanation…

    The explanation on the right is produced from this replay. Click any timestamp to jump to the evidence.

    An answer you can check in one click.

    An explanation nobody can verify is worthless on an incident call.

    Every claim carries a citation, and the citation is a jump link: you land on the moment, watch it, and decide for yourself. Where recording has a gap, the gap is stated rather than filled in with a guess.

    Analysis never sees masked text or media. Element labels come from accessible names, test ids, and roles.

    Grounding is enforced, not requested.

    A claim without a citation is rejected before it reaches you. An internal set of cases with known root causes gates every change we ship.

    Patient intake Demo
    Chose
    insurance-card.jpg 0:11
    Request
    POST 413 0:11
    Code
    upload.ts:88 0:13
    Cause
    No 10 MB check 0:13
    Confidence
    86%
    Same issue
    143 in 7 days
    Every claim cited

    Demo session. Claims from the explanation generated for it.

    Automatic, or on demand.

    Automatic where it is cheap. On demand where it is deep.

    • Summaries and clusters

      Sessions with a Signal get a triage pass and a summary. Clusters get a name and an explanation shared across every session in them. Daily digests tell you what changed since the last release.

      Automatic, included
    • Explain this moment

      Select a time range. Get what the user tried, what broke, the evidence of impact, the likely root cause with confidence, suggested fixes, and the linked backend trace.

      On demand, deep analysis
    • Repro steps and tests

      Numbered repro steps and a Playwright test using stable selectors, with masked data replaced by fixtures. Connect GitHub and Why drafts the PR.

      On demand, deep analysis
    • Chat with tools

      Ask in plain language. Why searches your sessions, gathers the evidence it needs, and answers with jump links.

      Ask Why
    • Natural-language search

      "Sessions where clinicians couldn't upload a document on iPad." The interpreted filter is shown, so you can trust the result list.

      Search
    • MCP server

      Any MCP client can pull a session's evidence, an error context, or repro steps straight into a coding session with a scoped token.

      Agents

    In the United States, under your ceiling.

    • It stays in the United States

      Analysis for a PHI tenant stays in the United States, with no path out of it, and a tripwire alerts on any attempt otherwise.

    • Depth where it earns it

      Triage, summaries, and cluster explanations are automatic and included. The deep work runs when a person asks for it. Economy mode trades a little speed for a lower bill when you would rather spend less than wait less.

    • A ceiling you set

      A monthly spend cap is a hard stop on metered analysis, and the included AI keeps running underneath it. Bring your own key on Business and Enterprise and the metering stops entirely.

    AI spend Monthly
    Triage
    Automatic, included
    Summaries
    Automatic, included
    Cluster explanations
    Automatic, included
    Deep analysis
    When a person asks
    Spend cap
    Hard stop on metered
    Your own key
    Metering stops
    PHI analysis
    Pinned to region
    Included AI keeps running

    Product updates by email.

    Leave a work email and we will tell you when something ships. Or book a demo with an engineer.

    Free for 1,000 stored sessions a month, with unlimited seats. Create a workspace and add one script tag.

    Start