The platform

Every inbound call, read and written down.

We transcribe each call and attach a structured read of what happened on it. Four more views are meant to sit on that record: the demand your phone is already capturing, the questions your team could not answer, one tracking record for every call and its revenue, and an API for when you would rather not open a dashboard.

14-day trial on Scale. No card, no contract, no setup fee. Two of the five parts below are still in build and the other three are not written yet — the table further down says which.

Five parts

What the platform does, part by part

We read from the systems that already carry your calls. We do not answer calls, route them, or hold your phone numbers.

Every call, not a sample

Every call gets a verdict.in build

We transcribe each recording, then attach a structured read: what the caller wanted, how ready they were to buy, what was promised, how it ended, and whether it was spam or voicemail. On all of them, not the handful somebody had time to listen to.

Call analysis in detail
Analyzed share · illustrative
1,271 of 1,284 calls got a verdict
  • What they wantedThe request in plain terms, plus the questions they asked — the raw material for demand and coaching.
  • Outcome and intentA closed vocabulary of outcomes plus an intent score, so you can filter instead of listening.
  • Estimated valueA range, not a false point estimate — anchored to what was actually discussed on the call.
  • Spam and voicemail flagsThe two categories that quietly inflate call counts and distort every average built on them.
  • Traceable labelsEvery label points at the part of the transcript that produced it. No black box.

What people actually called about

Hear what every caller wanted — including the things you don't sell yet.planned

Every call carries a request. Aggregated across hundreds of them, those requests are the clearest demand signal a business has, and most of it evaporates the moment the call ends.

Demand intelligence in detail
Requests, ranked · illustrative
  • Requests, rankedWhat callers asked for, grouped and counted, so a rising service shows up as a trend and not a hunch.
  • Unmet demandServices people asked for that you do not offer, or do not offer in that area — the expansion list your calls are already writing.
  • Shifts over timeWeek-on-week movement in what callers want, so seasonality and new competitors surface early.
  • Down to the callEvery number opens the calls behind it, with the sentence that produced the label.

Coach from evidence, not memory

Find the questions your team couldn't answer.planned

The most useful coaching material your business produces is the moment a caller asks something and nobody has a good answer. Those moments are invisible unless somebody listens to every call.

Team performance in detail
Unanswered, per week · illustrative
  • Unanswered questionsQuestions asked on calls that went unanswered, unclear, or got a promise to call back — ranked by how often they recur.
  • Per-person feedbackHow each person handles calls, with the specific calls behind the read. Objective input for a one-to-one, not a scoreboard.
  • Missed opportunitiesCalls that showed buying intent and ended without a next step.
  • Turn gaps into trainingRecurring gaps become a short list of what to teach next, with real caller wording to practise against.

One place for all of it

Keep track of everything, in one place you can audit.in build

Calls, sources, outcomes and revenue in one view — with the counting rules published, so the numbers survive an argument.

Tracking & revenue in detail
Answer rate · illustrative
  • Every call in one timelineAcross numbers, sources and locations, with outcome and value attached.
  • Joined to your marketingConnect Search Console and your call platform so you can see what a channel is genuinely worth.
  • Drills to raw rowsEvery aggregate opens the calls behind it. A number you cannot audit is a number you cannot argue with.
  • Published methodologyWe document exactly which field we treat as revenue and how duplicates are handled.

For when you don't want a dashboard

Ask your calls what happened.planned

A versioned REST API and a read-only MCP server, so your call data is available to your own tooling and to an AI session without anyone opening a chart.

API & MCP in detail
  • REST APIVersioned, documented, org-scoped keys with scopes and revocation.
  • MCP serverQuery calls, transcripts and revenue aggregates from Claude and other MCP clients.
  • Read-only by designThe MCP server queries your data. It cannot alter it.
  • Export alwaysFull CSV export on every tier — no gate, no support ticket.
The build ledger

What is built, and what is not

No customer logos, no counters, no accuracy figure. We launched in 2026 and have no customers yet, so here is the build state instead, part by part, and a dated changelog you can check it against.

in build — code exists, not self-serve planned — not written yet dated changelog →

PartStatusWhat that means if you sign up today
Call analysis in build Transcription and the outcome, intent, value and spam labels run. The connectors that feed them are not self-serve yet, so we set them up with you.
Demand intelligence planned Not built. There is no aggregated view of what callers asked for yet, and no unmet-demand list. The per-call record of what was requested is what it will be built from.
Team performance planned Not built. Unanswered-question detection and per-person feedback are design work, not shipped code. Do not buy on this yet.
Tracking & revenue in build Calls, sources, outcomes and revenue in one place, with the counting rules published. Working, not yet something you can turn on alone.
API & MCP planned Neither the REST API nor the MCP server exists today. When the MCP server ships it will be read-only. Several other vendors already ship one; we are not claiming to be first.

Connections in build

Ringba and Google Search Console. Usable with our help, not yet self-serve.

Connections planned

CallRail, Retreaver, CSV upload, Google Sheets, Slack digests and the MCP server. Each carries its own status on the integrations page.

Committed from day one

Full CSV export on every tier, month-to-month billing, a public dated changelog, and we never train models on your calls.

The labels are literal. Nothing gets marked live until you can connect it yourself and it syncs without us, and nothing gets a status for marketing reasons. Everything shipped so far is dated in the changelog. We hold no SOC 2 report; the security page says what we do and do not have.

Point it at one month of your own calls

Connect a source, let it read real calls, and check the labels against the recordings yourself. That is the only test of this that means anything. 14 days on Scale, no card.