Privacy-first AI transcription
Audio & video
to text. You control
what stays.
99%+ accuracy on audio and video. Speaker diarization, sentiment analysis, and NLP enrichment — all in one API call. HIPAA BAA-ready. Your data is never used to train models.
HIPAA BAA-Ready
SOC 2
TLS 1.2+ · AES-256
No Model Training
Built for
Capabilities
Everything you need.
Nothing you don't.
From raw audio to structured, enriched data — in a single API call, with a data lifecycle you define.
99%+ Accuracy
Powered by a custom Whisper-Large model, optimised for accented speech, noisy audio, and domain-specific vocabulary from legal, medical, and financial contexts.
Speaker Diarization
Identify who said what with word-level timestamps and confidence scores. Exports as JSON, SRT, VTT, TXT, or DOCX — ready for your downstream workflow.
Sentiment & NLP
Extract sentiment, named entities, key topics, concise summaries, and intent from audio — all in one pass. Ideal for call centre QA, research analysis, and conversation intelligence platforms.
Configurable Retention
Set how long your source files and transcripts are held — from 0 (purged immediately after processing) to 30 days. Your data lifecycle, under your control.
API-First
REST API with JSON responses, webhook callbacks for async workflows, and integrations with platforms like n8n. Upload via web if your team prefers a no-code interface.
99+ Languages
Broad multilingual coverage including accented and regional variants. Enterprise fine-tuning available for specialised vocabulary in legal, clinical, and financial domains.
How it works
Simple pipeline.
You set the rules.
Audio processed in-memory on Vocant's self-hosted servers. Your retention window is yours to configure — down to zero days.
01
Upload
Send your file
POST audio or video via the API, or upload via the web interface. MP3, MP4, WAV, MOV, FLAC and more accepted. Files are encrypted in transit (TLS 1.2+) and at rest (AES-256).
POST /v1/transcribe
Authorization: Bearer <key>
file=@hearing.mp4
diarize=true
sentiment=true
summarize=true
retention_days=0
02
Process
In-memory transcription + NLP
Files are processed in-memory on Vocant's self-hosted infrastructure. Whisper-Large transcribes with speaker labels and timestamps. Optional NLP enrichment runs in the same call.
← Whisper-Large (custom model)
← Speaker diarization: enabled
← Sentiment analysis: enabled
← Named entity recognition: on
← Summarization: enabled
← Training use: never
03
Retrieve & expire
Transcript delivered, source purged on schedule
Enriched transcript available instantly via API or dashboard. Source audio is purged per your retention setting. Raw audio is never included in system backups.
{
"status": "complete",
"retention_days": 0,
"training_use": false,
"sentiment": "neutral",
"summary": "Follow-up in 2 weeks; adjust dosage.",
"entities": [{"text": "Dr. Patel", "type": "PERSON"}],
"words": [...]
}
Privacy architecture
Transparent about exactly
how your data moves.
Data flow
File uploaded
Encrypted in transit via TLS 1.2+ and at rest via AES-256 on Vocant's self-hosted servers.
Processed in-memory
Audio is loaded and transcribed in-memory. NLP enrichment runs in the same pass. No training use.
Transcript stored in your account
Resulting transcript and metadata stored in your account for download. Accessible only to you.
Source file purged per your schedule
Raw audio deleted per your retention setting (0–30 days). Raw audio is never included in any system backups.
You control how long files live
Set source file retention from 0 to 30 days in your account profile. At zero, your file is purged immediately after processing completes.
Never trained on your data
Vocant does not use your audio, video, or transcripts to train generalized AI or machine learning models unless you expressly agree in writing.
Encrypted end to end
TLS 1.2+ in transit. AES-256 at rest. Staff access to your content is restricted and subject to documented escalation procedures.
No raw audio in backups
Raw customer audio is never included in system backups — a meaningful architectural distinction that supports enterprise and regulated procurement requirements.
HIPAA BAA on Enterprise
Business Associate Agreements are available on Enterprise plans to support ePHI workflows in regulated healthcare environments.
Use cases
Anywhere sensitive audio
demands discretion.
Legal
Depositions & proceedings
Transcribe legal audio without routing it through platforms subject to discovery. Enterprise plans include legal-trained models.
Healthcare
Clinical notes & consultations
HIPAA BAA-ready transcription of physician notes, patient consultations, and dictation. PHI retained only as long as you choose.
Finance
Earnings calls & board recordings
Transcribe MNPI-sensitive recordings without creating a third-party cloud copy. Set retention to zero for maximum control.
Journalism
Source interviews
Protect sources — audio purged on your schedule means no subpoenable copy sits on our infrastructure beyond what you allow.
HR & Compliance
Investigations & grievances
Transcribe sensitive HR proceedings without exposing employee data to external platforms. Export to your own systems only.
Contact Centres
Call QA & sentiment
Bulk-transcribe calls with sentiment, escalation signals, and named entities in one pass. Feed structured JSON into your QA platform.
Podcasts & Media
Production & captions
Fast, accurate transcripts with speaker labels for show notes, accessibility, and SEO. Exports to SRT, VTT, DOCX.
Research
Interviews & focus groups
Transcribe qualitative research audio under IRB-compliant data handling. Configure retention to match your ethics approval requirements.
API & Integrations
Add Vocant to what
you're already building.
A REST API that fits into existing products or new tools alike — no new infrastructure required. Works with n8n and automation platforms out of the box.
Extend existing products
Drop into your current stack
Add privacy-first transcription to apps you've already built. A single endpoint returns structured JSON.
- EHR & practice management platforms
- CRM and contact centre software
- Legal case management systems
- Meeting and collaboration tools
- Automation platforms (n8n, Zapier, Make)
Build something new
A foundation for voice-first tools
Build on Vocant from day one. The API returns transcripts, speaker labels, timestamps, sentiment, and entities.
- AI meeting assistants & notetakers
- Compliance monitoring dashboards
- Podcast & media production tools
- Clinical documentation workflows
- Voice-enabled search & discovery
Python
Node.js
cURL
# pip install vocant import vocant client = vocant.Client(api_key="vk_...") result = client.transcribe( file="call_recording.mp4", diarize=True, sentiment=True, entities=True, summarize=True, retention_days=0 # purge immediately after processing ) print(result.transcript) print(result.sentiment) # "neutral" | "positive" | "negative" print(result.summary) # concise abstractive summary of the transcript print(result.entities) # [{"text": "Dr. Patel", "type": "PERSON"}, ...] print(result.words) # word-level timestamps + speaker IDs
Compliance
Standards
enterprise teams
actually check.
enterprise teams
actually check.
HIPAA
BAA on Enterprise
SOC 2
Attested · Security & Privacy
GDPR
EU-compatible data handling
AES-256
TLS 1.2+ in transit
Pricing
Simple, transparent
pricing.
Every plan includes configurable retention (0–30 days), no model training, and full accuracy. Privacy is not a premium feature.
Free
$0
Great for trying Vocant
- 5 free transcriptions / day
- Whisper-based accuracy
- API access
- Web uploads
- Speaker diarization & timestamps
- Sentiment & NLP analysis
- Exports: SRT, VTT, TXT, DOCX
- Configurable retention (0–30 days)
Most popular
Pro
$29
per user / month
- 2,000 minutes / month
- Everything in Free +
- Faster processing
- Higher rate limits
- Priority support
- Webhook callbacks
Enterprise
Custom
SLA · BAA · SSO
- Unlimited transcription minutes
- HIPAA BAA + SSO / SAML
- Healthcare & legal-trained models
- Dedicated account management
- Priority SLA
- Private model fine-tuning (add-on)
- Security review & architecture docs
API access available on all plans. Healthcare and legal capabilities on Enterprise only.
Full pricing details →
FAQ
Common questions,
honest answers.
Audio and video files are processed in-memory on Vocant's self-hosted servers and immediately purged from the processing environment when transcription completes. Your resulting transcript is stored in your account until you delete it or close your account. Raw audio is never included in system backups.
No. Vocant does not use customer-submitted audio, transcripts, or other customer content to train generalized AI or machine learning models unless a customer expressly agrees in writing. Enterprise fine-tuning is always opt-in, isolated to your organization, and contractually governed.
Vocant is HIPAA BAA-ready. Business Associate Agreements are available on Enterprise plans. For regulated workflows, we recommend reviewing our privacy policy and contacting sales for a full compliance overview.
Yes. Vocant provides a REST API for file uploads, webhooks, and transcript retrieval. It integrates natively with automation platforms like n8n. Full documentation is available at app.vocant.ai/api-docs.html. API access is included on all plans.
Vocant exports as .srt, .vtt, .txt, and .docx — all with speaker labels and timestamps included. JSON is the native API response format, making it easy to feed into downstream workflows.
The free plan includes 5 transcriptions per day with no credit card required. All features including API access, diarization, sentiment analysis, and NLP are available on the free tier. Upgrade to Pro for 2,000 minutes/month and faster processing.
Start transcribing.
Leave no trace.
No credit card required. 5 free transcriptions per day. Your first API call is minutes away.