Livela sintesi · call analysis · resolving live
call analysis · hundreds to tens of thousands of calls a month

The calls alreadyhappened. whatnow?

La Sintesi turns every phone call into a transcript, an outcome, an agent score, and a metrics dashboard — under sixty seconds, behind a single API.

$0.05 / minuteunder 60s end-to-endAPI-first
NO CREDIT CARD · NO EXPIRATION
START FREE · $5 CREDITS

Create your workspace

No credit card. Credits never expire.

Encrypted · isolated per workspace · delete anytime

Already have an account? Sign in →

SAMPLE ▸
c84aSuccessful+0.824:12
c84bFollow-up+0.412:58
c84cUnsuccessful−0.301:47
c84dNo Contact0.000:38
c84eSuccessful+0.776:05
c84fNegative−0.550:52
c850Follow-up+0.363:21
c851Informational0.000:21
c852Successful+0.695:40
c84aSuccessful+0.824:12
c84bFollow-up+0.412:58
c84cUnsuccessful−0.301:47
c84dNo Contact0.000:38
c84eSuccessful+0.776:05
c84fNegative−0.550:52
c850Follow-up+0.363:21
c851Informational0.000:21
c852Successful+0.695:40
SENTIMENT TRACE · SAMPLE OUTPUT
+ POS− NEG
TRANSCRIBING· 62%
DECODING· 34%
SCORING·
INTAKEthe backlogconf 0.91
what high-volume call teams tell us — before the engine hears a thing

Call-center leaders know the shape of the problem. They just can't hear it.

01 · IN THEIR WORDS

We have hundreds of calls a day — there's no way to listen to them all.

02 · IN THEIR WORDS

We don't know which agents are performing and which aren't until it's too late.

03 · IN THEIR WORDS

We're manually classifying call outcomes in spreadsheets. It's a disaster.

QA DIRECTOR · CALL OPERATIONS

DECODEhow it resolvesconf 0.96
three stages, no black boxes · five layers of analysis

Three steps. No black boxes.Five layers of analysis.

Submit a recording. Wait under a minute. Get back a fully structured call record — the same one your team will query through the dashboard or your own tools.

01LAYER

Full transcript

Diarized · word-level timestamps · confidence

Speaker-labeled transcript with per-word confidence and offsets. Agent and customer turns are separated automatically.

7D TREND
02LAYER

Outcome classification

6 default categories · fully editable per workspace

Every call is bucketed into a business outcome — Successful, Follow-up, Unsuccessful, Negative, No Contact, Informational. Or define your own taxonomy.

7D TREND
03LAYER

Factual call summary

Opening · discussion · concerns · outcome

A short, factual recap of what actually happened on the call. No fluff. No invented context.

7D TREND
04LAYER

Agent performance score

1–5 · coaching remarks per call

One strength, one improvement area — two improvements when the score is three or below. Built for QA review, not vanity metrics.

7D TREND
05LAYER

Rich call metrics

Talk time · silence · interruptions · WPM

Per-speaker numbers you can act on. Roll them up across an agent, a team, or a whole operation.

7D TREND
06 · OUTPUT

One JSON. Everything.

Transcript, outcome, summary, score, and metrics — delivered as one call record over REST. No glue code, no scraping.

See the response shape
SAMPLE CALL RECORD · call_c84aEXAMPLE OUTPUT
OUTCOME
Successful Outcome
AGENT SCORE
4 / 5
SENTIMENT
DURATION
4:12 · 5 min billed
SUMMARY

Customer called to renew an expiring policy. The agent confirmed coverage, answered two billing questions, and processed a 12-month renewal. No escalation needed.

COACHING · STRENGTH

Clear, unprompted pricing breakdown.

COACHING · IMPROVE

Confirm the callback number earlier.

TALK A/C
46 / 54%
SILENCE
8%
INTERRUPTS
1
WPM
148
CLASSIFYyour categoriesconf 0.93
ships with a 6-category default — replace every one

Your categories.Not ours.

The default 6-category taxonomy is an industry-neutral starting point. Rewrite it for your business — inbound support, sales, fundraising, real-estate leads, debt collection — and the scoring logic adapts to your taxonomy automatically.

FULLY EDITABLE PER WORKSPACE
DEFAULT · taxonomy-v1EDITABLE
  • 01Successful Outcome
    edit →
  • 02Follow-up Scheduled
    edit →
  • 03Unsuccessful Outcome
    edit →
  • 04Negative Experience
    edit →
  • 05No Contact
    edit →
  • 06Informational
    edit →
EXTRACTthe metersconf 0.95
the numbers that move a call operation

The numbers that movea call operation.

TALK TIME

Per-speaker minutes and percentage of the call. See who actually drove the conversation.

SILENCE RATIO

Dead air, hold time, and pauses longer than 1.5s — tallied and surfaced per call.

INTERRUPTIONS

Counted per side. Agent over customer, and vice versa. A small number with a big QA story.

WORDS / MIN

Speaking pace per speaker — and how it shifts mid-call when the conversation turns.

WEIGHthe alternativesconf 0.95
three ways to turn calls into structured data

Build it, buy a platform,or send the first call.

Most teams weigh the same three options. Here's the honest trade-off — no straw men.

BUILD IT IN-HOUSE
  • Stitch together speech-to-text, an LLM, storage, and scoring prompts
  • Weeks of engineering before the first usable result
  • You own the evals, the taxonomy design, and the upkeep
LEGACY CALL-QA SUITES
  • Book a demo, sit through a sales call
  • Annual contract and onboarding; pricing on request
  • Built for enterprise procurement, not a self-serve trial
MOST DIRECT
LA SINTESI
  • Sign up free — $5 in credits, no credit card
  • First structured record back in under 60 seconds
  • $0.05/minute, no contract, one REST API

NO SALES CALL · NO CONTRACT · LIVE IN UNDER A MINUTE

METERthe rateconf 1.00
$0.05 / minute · pre-paid credits · cached re-views are free

That's it.$0.05 a minute. 2 free re-analyses per call.

$0.05 / MIN · CREDITS NEVER EXPIRE

Billed per started minute — a 3:01 call counts as 4 minutes, same as every telecom you've ever used. Calls under a minute count as one. Re-viewing analysis is always free; re-running the AI on a call is free for the first 2 attempts, then $0.03 flat each.

These are prepaid credit packs, not subscription tiers — same $0.05/minute on every one. No seats, no minimums, no contract. Pick the balance that fits your volume and top up when it runs low; credits never expire.

STARTER
$10~200 minutes

Kick the tires on real calls. Top up whenever you're ready.

BUY CREDITS
STANDARD
$25~500 minutes

A bigger balance for a small team running calls every day.

BUY CREDITS
MOST POPULAR
OPERATOR
$50~1,000 minutes

Most teams land here — about a month of steady volume in one top-up.

BUY CREDITS
VOLUME
$100~2,000 minutes

For dialers and call centers running tens of thousands of calls a month. Email hello@lasintesi.co for bulk rates (up to ~40% off).

BUY CREDITS
~40%
off, for volume.

Running thousands of minutes a month? Email hello@lasintesi.co with your monthly volume and we'll send a quote — typically up to ~40% off the list rate for committed volume.

CONTACT SALES
ESTIMATE · DRAG TO ADJUST
1,000 minutes / mo200 calls
$50.00/ month
NO CREDIT CARD · NO EXPIRY · 2 FREE RE-ANALYSES / CALLSTART FREE · $5 CREDITS →
EXPORTintelligence outconf 0.97
one REST call in, one structured record out
FOR DEVELOPERS

API-first.Always was.

Every value the dashboard renders is reachable through one REST API. Submit a call, poll or fetch the structured record back, and push it straight into your CRM, BI tool, or internal reporting stack.

  • REST API
    Key-based auth · per-key rate limiting · OpenAPI 3.0 spec
  • One structured record
    Transcript · outcome · summary · score · metrics
  • Cached responses
    Already-analyzed calls return instantly at zero cost
// Submit a recording — get the full analysis under 60s.
const res = await fetch("https://app.lasintesi.co/api/analyze", {
  method: "POST",
  headers: {
    "x-api-key": process.env.LS_API_KEY!,
    "Content-Type": "application/json",
  },
  body: JSON.stringify({
    callUrl: "https://cdn.acme.co/calls/0421.mp3",
    agentId: "agent_834",
  }),
});

const { callRecords, status } = await res.json();

console.log(callRecords[0].sentiment); // "Successful Outcome"
console.log(callRecords[0].summary);   // factual recap
console.log(status);                   // "completed"
VERIFYsoundcheckconf 0.94
questions, honestly answered

Questions,honestly answered.

Couldn't find it? Email hello@lasintesi.co and we'll get back to you.

01How long does an analysis take?
Most calls finish in under 60 seconds end-to-end — transcript, outcome, summary, score, and metrics. Long calls scale roughly with duration. Results land in the dashboard and become available through the API the moment they're ready.
02What audio formats do you accept?
MP3, WAV, M4A, and FLAC up to 100 MB. Upload the file directly through the dashboard or the API, or submit a signed URL we can fetch from your own storage.
03Can I change the outcome categories?
Yes. The 6 defaults are a starting point — every workspace can rewrite the taxonomy in the Analysis configuration. We compile your categories into the scoring prompt automatically.
04Where is my audio stored?
Audio files are stored in private, encrypted cloud object storage that only your workspace can read. We never share recordings with other tenants, and you can delete a call — and its analysis — at any time.
05Who owns the analysis?
You do. Transcripts, outcomes, scores, and metrics belong to your workspace and are reachable through the dashboard or the REST API. We never resell or repurpose your call data.
06How does the free trial work?
$5 of free credits land in your wallet the moment you sign up, no credit card. After that you top up in $10/$25/$50/$100 increments. Re-viewing already-analyzed calls is always free — re-running the AI analysis on a call has its own quota (see below).
07How is a call billed if it's not a round number of minutes?
Partial minutes round up to the next whole minute — standard telecom-style billing. So a 1:20 call counts as 2 minutes ($0.10 at the default rate) and a 5:01 call counts as 6 ($0.30). Calls under a minute count as one minute. The exact rounded minute count and unit price always appear in your transaction history.
08What about re-analyzing a call?
Re-viewing existing analysis is always free. Re-running the AI on the same transcript is free for the first 2 attempts per call, then $0.03 flat per re-analysis — regardless of call length. Each charge shows in your transaction history with the attempt number.
09What languages do you support?
Multilingual transcription is supported. Outcome and scoring quality is best on languages with strong training coverage. If you have a specific language requirement, contact us before you commit.
10Can I get the data out programmatically?
Yes. Every value the dashboard renders — transcripts, outcomes, summaries, scores, and metrics — is reachable through the REST API. We publish an OpenAPI spec so you can generate a client in whichever language you use. API keys activate after your first top-up — the free signup credit lets you analyze calls in the dashboard right away, and a single top-up unlocks programmatic access.
RESOLVEsend the first oneconf 0.98
NO CREDIT CARD · NO EXPIRATION

$5 in credits.On us.

Spin up a workspace, drop in your first recording, and watch the transcript, outcome, and score land in your dashboard before your coffee's cold.