Skip to main content

Evaluate · provider-neutral voice comparison

Compare voice outputs

Compare the same scenario, listen before judging, and reveal the evidence behind each result.

Start with what you hear. Provider identity, measurements, methodology, and sanitized technical evidence remain available when you need them.

Read the versioned methodology
Why one comparison is not a universal ranking

Voice results depend on the text, voice, model, configuration, environment, and measurement method. ONE preserves those conditions so a result can be understood without turning one run into a universal claim.

Evaluation detail

This changes presentation only. It never changes access, ownership, provider policy, or what an action can do.

Saved on this device.

  • DefaultFixture evidence
  • ProvidersCapability-driven
  • Live callsProtected + confirmed
  • PrivacyPrivate + ephemeral
Execution evidence

Standardized mode aligns the exact text and comparable PCM/WAV output boundary. It does not claim provider controls are equivalent.

01 · Scenario

Choose one canonical script

Versioned preset

The exact text and its content hash become part of the private evidence bundle.

186 / 600

Use non-sensitive text only—no personal, confidential, or regulated content. Phase 1 runs are private and ephemeral; scripts, audio, and ratings are not sent to analytics.

Preset method note

Listen for clarity, warmth, alphanumeric pronunciation, and a natural question ending.

02 · Voices

Select two to four providers

0 selected

Loading provider readiness without making provider calls…

03 · Review

Confirm the evidence boundary

No provider spend. Fixture mode replays deterministic evidence and does not contact provider APIs.

Fixture mode is ready. Nothing runs until you confirm.

Select between two and four available providers.

Full evidence begins after one explicit run

Each provider will finish independently. A timeout or failure never erases another provider’s successful audio.

Evidence portability

Private by default. Portable by choice.

Export sanitized configuration and provenance, or inspect a compatible bundle without rerunning provider calls. Audio is not embedded.

Open benchmark planning and leaderboard evidencePlan fixture benchmarks, inspect ranking eligibility, and review integrity evidence.

Canonical benchmark preview

Verifiable results, without a composite winner

Explore a deterministic, synthetic leaderboard fixture. It is nonbillable, makes no provider-quality claim, and is not eligible for public evidence or publication.

Fixture-onlyNonbillableSynthetic evidenceNon-public previewone-benchmark/1.0.0one-benchmark-methodology/1.0.0
Open methodology

Ready. This preview makes no provider call.

Benchmark setup

Validate one canonical fixture observation

Choose two to four exact fixture lanes. ONE validates the plan through the shared action runtime, then materializes one canonical private result locally.

Nonbillable · zero provider calls
Providers and exact fixture models

Fixture comparability is available for every listed lane; adapter readiness is disclosed separately.

Idle: Ready to validate a nonbillable fixture benchmark.

Filter this fixture snapshot

Only dimensions represented by this deterministic fixture are available.

Objective measurementsDisplayed below with metric provenance and sample statistics.
Human judgmentsKept separate. This synthetic preview contains no human ratings.
Automated judgmentsReserved and isolated. No model-judged evidence appears here.

Exact provider

Cartesia

Tie · #1

Exact model

fixture-cartesia-tts-v1

Objective evidenceLocal deterministic fixtureRanked
Metric
fixture-interaction-success · 1 boolean
Sample count
n=1
Median
Unavailable · needs n≥3
p95
Unavailable · needs n≥20
Freshness
Fixture only · 2026-08-27 00:00:00 UTC
Public eligibility
Not eligible
Visibility
Private
Publication
Draft
Integrity
Unsigned
Signature status
Unsigned
Suite
one-fixture-preview-suite v1.0.0
Case
benchmark-fixture v1.0.0
Input hash
sha256:9f2e37000331808e54bd06f1cf5171922568bfee2de6fc47c56a38eabfafd8b3
Exact voice
fixture-cartesia-voice-v1
Configuration hash
sha256:18ca70f0e7361d9d63897e69cdae07b386d6df994748ce169d2200d8db2b26ab
Provider readiness
adapter-backed
Adapter version
one-deterministic-fixture/1.0.0
Model version
one-deterministic-fixture/1.0.0
Sponsorship
None disclosed for this evidence
Runtime environment
fixture
Region
Unknown / not applicable
Transport
local-memory
Codec
audio/wav · 24000 Hz · 1 channel
Exclusions
1
Why ranked here?

Methodology: Identical-script TTS (one-tts-identical-script v1.0.0). Profile fixture-preview-interaction-success v1.0.0 compares mean for fixture-interaction-success; Higher is better.

Comparable input: suite one-fixture-preview-suite v1.0.0; case benchmark-fixture v1.0.0; input sha256:9f2e37000331808e54bd06f1cf5171922568bfee2de6fc47c56a38eabfafd8b3.

Measurement provenance: fixture; one-server; not-applicable; synthetic; method one-deterministic-fixture-preview; source schema one-fixture-preview-metric/1.0.0.

Runtime: fixture; local-deterministic-fixture; local-memory; audio/wav · 24000 Hz · 1 channel.

Sample count: n=1; this profile requires n≥1. Median and p95 are shown only when their own minimum sample requirements are met.

Evidence: Objective evidence synthetic fixture measurements; no human or automated judgments are merged into this rank.

Sponsorship: None disclosed for this evidence. Sponsorship never changes rank or defaults.

Freshness: Fixture only as of 2026-08-27 00:00:00 UTC.

Integrity: Unsigned under one-canonical-json/1.0.0.

Visibility, publication, and signature: Private; Draft; Unsigned.

Digest: Unavailable for this fixture snapshot

Exclusions:
  • Synthetic fixture evidence is not public benchmark evidence.
Limitations:
  • This is a deterministic fixture-only preview. No provider was called and no provider quality, latency, availability, or model-support claim is made.
  • Every listed provider receives the same synthetic value and tied rank solely to exercise the generic leaderboard path.

No composite score is produced. Equal fixture measurements remain tied and do not imply provider quality.

Exact provider

Deepgram

Tie · #1

Exact model

fixture-deepgram-tts-v1

Objective evidenceLocal deterministic fixtureRanked
Metric
fixture-interaction-success · 1 boolean
Sample count
n=1
Median
Unavailable · needs n≥3
p95
Unavailable · needs n≥20
Freshness
Fixture only · 2026-08-27 00:00:00 UTC
Public eligibility
Not eligible
Visibility
Private
Publication
Draft
Integrity
Unsigned
Signature status
Unsigned
Suite
one-fixture-preview-suite v1.0.0
Case
benchmark-fixture v1.0.0
Input hash
sha256:9f2e37000331808e54bd06f1cf5171922568bfee2de6fc47c56a38eabfafd8b3
Exact voice
fixture-deepgram-voice-v1
Configuration hash
sha256:18ca70f0e7361d9d63897e69cdae07b386d6df994748ce169d2200d8db2b26ab
Provider readiness
adapter-backed
Adapter version
one-deterministic-fixture/1.0.0
Model version
one-deterministic-fixture/1.0.0
Sponsorship
None disclosed for this evidence
Runtime environment
fixture
Region
Unknown / not applicable
Transport
local-memory
Codec
audio/wav · 24000 Hz · 1 channel
Exclusions
1
Why ranked here?

Methodology: Identical-script TTS (one-tts-identical-script v1.0.0). Profile fixture-preview-interaction-success v1.0.0 compares mean for fixture-interaction-success; Higher is better.

Comparable input: suite one-fixture-preview-suite v1.0.0; case benchmark-fixture v1.0.0; input sha256:9f2e37000331808e54bd06f1cf5171922568bfee2de6fc47c56a38eabfafd8b3.

Measurement provenance: fixture; one-server; not-applicable; synthetic; method one-deterministic-fixture-preview; source schema one-fixture-preview-metric/1.0.0.

Runtime: fixture; local-deterministic-fixture; local-memory; audio/wav · 24000 Hz · 1 channel.

Sample count: n=1; this profile requires n≥1. Median and p95 are shown only when their own minimum sample requirements are met.

Evidence: Objective evidence synthetic fixture measurements; no human or automated judgments are merged into this rank.

Sponsorship: None disclosed for this evidence. Sponsorship never changes rank or defaults.

Freshness: Fixture only as of 2026-08-27 00:00:00 UTC.

Integrity: Unsigned under one-canonical-json/1.0.0.

Visibility, publication, and signature: Private; Draft; Unsigned.

Digest: Unavailable for this fixture snapshot

Exclusions:
  • Synthetic fixture evidence is not public benchmark evidence.
Limitations:
  • This is a deterministic fixture-only preview. No provider was called and no provider quality, latency, availability, or model-support claim is made.
  • Every listed provider receives the same synthetic value and tied rank solely to exercise the generic leaderboard path.

No composite score is produced. Equal fixture measurements remain tied and do not imply provider quality.

Exact provider

ElevenLabs

Tie · #1

Exact model

fixture-elevenlabs-tts-v1

Objective evidenceLocal deterministic fixtureRanked
Metric
fixture-interaction-success · 1 boolean
Sample count
n=1
Median
Unavailable · needs n≥3
p95
Unavailable · needs n≥20
Freshness
Fixture only · 2026-08-27 00:00:00 UTC
Public eligibility
Not eligible
Visibility
Private
Publication
Draft
Integrity
Unsigned
Signature status
Unsigned
Suite
one-fixture-preview-suite v1.0.0
Case
benchmark-fixture v1.0.0
Input hash
sha256:9f2e37000331808e54bd06f1cf5171922568bfee2de6fc47c56a38eabfafd8b3
Exact voice
fixture-elevenlabs-voice-v1
Configuration hash
sha256:18ca70f0e7361d9d63897e69cdae07b386d6df994748ce169d2200d8db2b26ab
Provider readiness
adapter-backed
Adapter version
one-deterministic-fixture/1.0.0
Model version
one-deterministic-fixture/1.0.0
Sponsorship
None disclosed for this evidence
Runtime environment
fixture
Region
Unknown / not applicable
Transport
local-memory
Codec
audio/wav · 24000 Hz · 1 channel
Exclusions
1
Why ranked here?

Methodology: Identical-script TTS (one-tts-identical-script v1.0.0). Profile fixture-preview-interaction-success v1.0.0 compares mean for fixture-interaction-success; Higher is better.

Comparable input: suite one-fixture-preview-suite v1.0.0; case benchmark-fixture v1.0.0; input sha256:9f2e37000331808e54bd06f1cf5171922568bfee2de6fc47c56a38eabfafd8b3.

Measurement provenance: fixture; one-server; not-applicable; synthetic; method one-deterministic-fixture-preview; source schema one-fixture-preview-metric/1.0.0.

Runtime: fixture; local-deterministic-fixture; local-memory; audio/wav · 24000 Hz · 1 channel.

Sample count: n=1; this profile requires n≥1. Median and p95 are shown only when their own minimum sample requirements are met.

Evidence: Objective evidence synthetic fixture measurements; no human or automated judgments are merged into this rank.

Sponsorship: None disclosed for this evidence. Sponsorship never changes rank or defaults.

Freshness: Fixture only as of 2026-08-27 00:00:00 UTC.

Integrity: Unsigned under one-canonical-json/1.0.0.

Visibility, publication, and signature: Private; Draft; Unsigned.

Digest: Unavailable for this fixture snapshot

Exclusions:
  • Synthetic fixture evidence is not public benchmark evidence.
Limitations:
  • This is a deterministic fixture-only preview. No provider was called and no provider quality, latency, availability, or model-support claim is made.
  • Every listed provider receives the same synthetic value and tied rank solely to exercise the generic leaderboard path.

No composite score is produced. Equal fixture measurements remain tied and do not imply provider quality.

Exact provider

Fish Audio

Tie · #1

Exact model

fixture-fish-audio-tts-v1

Objective evidenceLocal deterministic fixtureRanked
Metric
fixture-interaction-success · 1 boolean
Sample count
n=1
Median
Unavailable · needs n≥3
p95
Unavailable · needs n≥20
Freshness
Fixture only · 2026-08-27 00:00:00 UTC
Public eligibility
Not eligible
Visibility
Private
Publication
Draft
Integrity
Unsigned
Signature status
Unsigned
Suite
one-fixture-preview-suite v1.0.0
Case
benchmark-fixture v1.0.0
Input hash
sha256:9f2e37000331808e54bd06f1cf5171922568bfee2de6fc47c56a38eabfafd8b3
Exact voice
fixture-fish-audio-voice-v1
Configuration hash
sha256:18ca70f0e7361d9d63897e69cdae07b386d6df994748ce169d2200d8db2b26ab
Provider readiness
adapter-backed
Adapter version
one-deterministic-fixture/1.0.0
Model version
one-deterministic-fixture/1.0.0
Sponsorship
None disclosed for this evidence
Runtime environment
fixture
Region
Unknown / not applicable
Transport
local-memory
Codec
audio/wav · 24000 Hz · 1 channel
Exclusions
1
Why ranked here?

Methodology: Identical-script TTS (one-tts-identical-script v1.0.0). Profile fixture-preview-interaction-success v1.0.0 compares mean for fixture-interaction-success; Higher is better.

Comparable input: suite one-fixture-preview-suite v1.0.0; case benchmark-fixture v1.0.0; input sha256:9f2e37000331808e54bd06f1cf5171922568bfee2de6fc47c56a38eabfafd8b3.

Measurement provenance: fixture; one-server; not-applicable; synthetic; method one-deterministic-fixture-preview; source schema one-fixture-preview-metric/1.0.0.

Runtime: fixture; local-deterministic-fixture; local-memory; audio/wav · 24000 Hz · 1 channel.

Sample count: n=1; this profile requires n≥1. Median and p95 are shown only when their own minimum sample requirements are met.

Evidence: Objective evidence synthetic fixture measurements; no human or automated judgments are merged into this rank.

Sponsorship: None disclosed for this evidence. Sponsorship never changes rank or defaults.

Freshness: Fixture only as of 2026-08-27 00:00:00 UTC.

Integrity: Unsigned under one-canonical-json/1.0.0.

Visibility, publication, and signature: Private; Draft; Unsigned.

Digest: Unavailable for this fixture snapshot

Exclusions:
  • Synthetic fixture evidence is not public benchmark evidence.
Limitations:
  • This is a deterministic fixture-only preview. No provider was called and no provider quality, latency, availability, or model-support claim is made.
  • Every listed provider receives the same synthetic value and tied rank solely to exercise the generic leaderboard path.

No composite score is produced. Equal fixture measurements remain tied and do not imply provider quality.

Telephony lab · simulation onlyNo live call capability

ONE Voice Concierge

What would you like to accomplish?

Describe an outcome. ONE will explain up to three existing journeys, then wait for your choice.
Choose the detail for this guidance

This changes presentation only. It never changes access, ownership, provider policy, or what an action can do.

Saved on this device.

Why this view?

Add short explanations where a concept becomes useful. The destination IDs, permissions, costs, and outcome semantics never change with this setting.

Guide only: no action, provider, upload, content save, spending, or scenario run can happen here. Your existing presentation-depth preference may persist.