Customer-hosted · your AWS account
Conversations become work.
Capture the conversations that matter. EchoWorks turns them into searchable knowledge, actions, records, and intelligence — inside your environment.
Captures from
- Google Meet
- Google Calendar
- Zoomcoming
- Microsoft Teamscoming
Google Calendar and Google Meet are the first native acquisition track in active development. When a native transcript isn't available, the EchoWorks Recorder provides local fallback capture — it is not the primary way meetings are captured.
How it works
Signal
Approved scheduled meetings are discovered from your calendar, and native provider transcripts are acquired directly — no meeting bot required when a native artifact is available.
Understanding
Transcript and source material become evidence-backed summaries, decisions, actions, and commitments — deterministic facts computed first, the model interpreting significance second.
Work
Reviewed results become searchable records, publish to systems like HubSpot, and stay accessible to approved AI tooling through MCP.
Your data plane stays in your account.
Recordings, transcripts, and derived intelligence are written to your own AWS account. EchoWorks ships the control plane, entitlements, and signed discovery — never your conversation content.
Your AWS account holds
- Provider credentials and configuration
- Meeting and transcript content
- Derived intelligence
- Identity and RBAC
- Integration credentials
- Retention policy and enforcement
- Your model/inference usage
EchoWorks vendor infrastructure holds
- Commercial and technical entitlement
- Signed license tokens
- Deployment and discovery metadata
- Release artifacts
- Optional content-free health metadata
The vendor control plane does not need — and does not receive — transcript content, audio, derived meeting intelligence, or CRM content.
Platform direction
Six things every conversation should produce.
Not a transcript in a folder. A record your systems can act on, with the evidence attached. Capabilities below describe the EchoWorks platform direction; they roll out incrementally, starting with Google Calendar and Meet.
Traceable summaries
Every generated claim links back to the transcript moment it came from, so a summary survives being questioned.
Owned actions
Commitments become assigned tasks with an owner and due date, synced to HubSpot — with more destinations planned.
Decision records
Decisions are captured separately from discussion, with participants and the alternatives considered.
Searchable knowledge
Ask across every conversation your permissions allow and get answers with sources, not vibes.
Integration and sync
Field-level mapping into HubSpot in both directions, with a per-connector audit of what was written and when.
Consent and retention
Per-jurisdiction consent capture, redaction of sensitive spans, and retention windows you control.
Built around
What EchoWorks is built around
Meeting content, transcripts, and derived intelligence are stored and processed entirely inside your AWS account.
Signed, KMS-backed license tokens are verified locally by your deployment — no runtime phone-home for ordinary operation.
EchoWorks acquires native meeting transcripts first. The Recorder is a fallback path, not the primary capture mechanism.
Every generated summary, decision, and action is built to carry a citation back to the segment it came from.
EchoWorks is in active development and early access. We publish measured outcomes once we have them — not before.
Pricing
Pricing built around your deployment and usage.
Infrastructure runs in your AWS account, so you pay EchoWorks for the control plane and support — not for infrastructure markup. Final pricing is set with each customer during the early-access program.
For teams piloting EchoWorks against real meetings.
- Google Calendar + Meet acquisition pilot
- Conversation intelligence capabilities as they become available
- Direct access to the team building it
For organizations moving from pilot to production.
- Platform capabilities as they reach general availability
- SSO and provisioning (planned)
- Dedicated onboarding
Full customer-hosted deployment in your AWS account, your keys.
- Deployed in your AWS account under your security and retention controls
- Direct control over retention windows and consent policy
- Model routing you control, including Bedrock
What buyers ask first.
Where does meeting data live?
In your AWS account. Recordings, transcripts, and derived records are written to storage such as S3 and DynamoDB inside your own environment — never a vendor-hosted database.
Does EchoWorks send transcripts to its own SaaS account?
No. The vendor control plane holds customer identity, commercial entitlement, signed licenses, and deployment metadata — it never receives conversation content.
How does native transcript acquisition work?
The first integration is being built around Google Calendar discovery and native Google Meet transcripts, without a bot when the native artifact is available.
What happens when a native transcript is unavailable?
Acquisition falls back in order: platform recording, then the EchoWorks Recorder, then manual upload. Every record keeps its source and provenance.
Can we use our own AWS account and models?
Yes. EchoWorks deploys into your AWS account, cloud and model usage is billed to you directly, and each processing stage can be routed to the provider your team has approved.
How does licensing work?
Licenses are signed, KMS-backed tokens that your deployment verifies locally — no vendor phone-home for ordinary operation. A contract change issues a new token; the prior one is retained in history.
Which meeting providers are supported or planned?
Google Calendar and Google Meet are the first native acquisition track, in active development. Zoom and Microsoft Teams native adapters are planned next.
Does EchoWorks require a bot in every meeting?
No. Native transcript acquisition is preferred wherever a provider supports it. When it isn't available, the EchoWorks Recorder — a local capture process on the approved user's own endpoint — provides fallback capture. It does not join the meeting as a bot.
How does HubSpot integration fit?
Planned HubSpot integration is designed to publish reviewed summaries, decisions, and actions to associated HubSpot records, without copying the full transcript by default.
How are generated claims tied back to evidence?
The intelligence model is designed so every generated claim retains a reference to the transcript segment it came from — nothing is asserted without a resolvable source.
See it against your own meetings.
Join the early-access program to run EchoWorks against a real set of meetings in your own AWS account.