Everything Transcriwise does

01

Transcription

From raw audio to text, on the right engine for each recording

The core of the product: turn audio or video into faithful text, auto-matching (or hand-picking) the engine best suited to the file, language, and context.

Transcription types

Four contexts with their own settings: General (class, interview, lecture), Legal Practice, Meeting, and Subtitles.

Accepted inputs

File upload (MP3, WAV, M4A, MP4, MOV, MKV, TXT, MD, PDF, DOCX), public URLs, and cloud import (Dropbox, Gmail, Google Drive).

Audio and output language

Recognition in up to ~100 languages depending on the provider; the formatted document can be delivered in another language (the formatting AI translates the result).

Diarization

Speaker separation: each turn is attributed to a speaker. Name and role labels can be reviewed in the editor.

Expected speakers

Provide an estimated count and a roster of participants (with roles) to guide separation and labeling.

Narrative voiceapostila

Adjusts the text to the teacher/speaker voice when the goal is study material, without losing fidelity.

02

ASR providers

Nine engine families, a choice separate from the integration

Which app you connect and which engine transcribes the audio are different decisions. Each family has distinct strengths by language, mode, and cost — the system routes by context or you pin it manually.

Whisper

Large V3 and V3 Turbo models (via serverless GPU), with long files and configurable diarization.

Deepgram

Nova-3, Deepgram's newest, highest-accuracy model, with smart formatting (punctuation, paragraphs, speaker turns).

AssemblyAI

Universal models with diarization, timestamps, custom vocabulary, and optional analysis.

ElevenLabs

Scribe models, with audio events, diarization, and custom vocabulary.

Soniox

Multilingual transcription with language identification and translation in compatible flows.

Gladia

File and live, with diarization, code-switching, and custom vocabulary.

Voxtral

Mistral's proprietary model, tuned for difficult audio.

iFlytek

Specialist in Mandarin and Cantonese; Basic (18 languages) and LLM (30+ languages) engines.

Alibaba Cloud

Qwen-ASR, Paraformer, and SenseVoice models for different scenarios.

03

Live transcription

Talk, and watch the text appear

Transcribe meetings, hearings, and interviews as they happen, straight from your microphone or a browser tab.

Real-time capture

Microphone or browser-tab audio, transcribed live.

Speaker panels

Speakers appear and update in real time during the session.

Subtitle overlay

Configurable subtitle overlay during capture.

Active AI sidebar

The AI chat stays available mid-session to ask and summarize as you speak.

Exportable session

When you stop, it all saves as a transcript, ready to format and export.

04

Formatting and documents

The text becomes a document — in the format the work needs

After transcription, the pipeline structures the content: model, style, tables, and review level to deliver a finished document without losing the link to the audio.

Formatting mode

Sets how the text is treated — from didactic study guide to strict fidelity, meeting minutes, or a legal document.

Processing presets

Shortcuts (Fast, Standard, Professional, Premium) combining base model, execution profile, and audit level; any manual tweak makes it Custom.

Document style

DOCX look — from the pipeline default to Classic, Minimal, Executive, Academic, or Legal compact.

RAW + formatted

You get the raw transcript (RAW) and the formatted document side by side — the raw base stays accessible.

Pro tables

Generates the summary tables with the advanced model while the body runs on the economy model.

Custom prompt

Customize extras/tables or, in full-style mode, replace the style rules — without exposing the internal prompt.

RAW cache

Reuses the raw transcript already generated for the same file, speeding reprocessing and cutting cost.

05

Fidelity audit

Trust in what got written

A review layer that compares the document with the reference transcript and flags where the AI may have drifted from the audio.

Audit levels

Off, Fidelity, or Full — the higher the level, the stricter the check against the audio.

Divergence flags

Flags possible insertions, omissions, and low-confidence passages for targeted review.

Score per segment

Fidelity score per passage and per speaker, with an exportable report.

Assisted corrections

Optional auto-apply of structural (default) and content corrections identified by the audit.

Reasoning and execution

Reasoning level and execution profile control analysis depth and the balance of speed, cost, and rigor.

06

Study material

From the class to study material

Optional layer that turns the transcript into a study guide: summaries, maps, exercises, and per-topic reinforcements — enabled only when it helps.

Executive summary

An actionable synthesis at the top of the material.

Content mind map

A diagram of concepts and relationships, with a configurable number of maps.

Practice exercises

Multiple choice, true/false, and essay, with the amount you choose (or the AI decides).

Per-topic reinforcements

Mini mind map and quick quiz at the end of each topic.

Diagrams and illustrations

Mermaid flowcharts and AI illustrations (high fidelity, concepts, and flowcharts).

Glossary and illustrated guide

A glossary of technical terms and an illustrated guide that pairs slides (attached or extracted from video) with the transcript.

07

Legal Practice

Hearings and proceedings, with the right vocabulary and form

A context built for law, with procedural fields, forensic vocabulary, and jurisdiction-specific templates.

Case fields

Jurisdiction, hearing type/branch, case number, court, district/city, date, and legal objective.

Legal modes

Formatting by document type — hearing, testimony, appellate decision, and notarial document — each with its own structure.

Formal minutes

Structured minutes, with options for indirect speech, summarized minutes, and timestamps in the result.

Hearing participants

Register parties by procedural role (judge, attorney, prosecutor, witness…) to guide diarization.

75 specialized templates

Structures by country, court, and proceeding type — civil, labor, criminal, small claims, appellate, and more.

Legal dictionary and formatting

Legal corrections and normalized acronyms applied automatically, with DOCX styles and a case header.

08

Studio

From transcript to content, in one click

Tools that turn one or more transcripts into finished formats. Pick the sources, choose the output, and Studio generates it — all editable afterward in the Canvas.

Podcast / Audio

An audio conversation between two hosts, with quality presets.

Summary and briefing

A direct synthesis and an executive briefing with decisions, risks, and next steps.

Study guide and flashcards

Sectioned lessons with objectives and self-scoring cards (know / review / don’t know).

Mind map and FAQ

A visual concept diagram and Q&A generated from the content.

Review table and timeline

Compare sources and evidence in a table; order events and milestones on a timeline.

Document Canvas

An editor with revisions, version history, and an image gallery.

09

AI Sidebar

An analysis session beside the document

Chat with any transcript and pull out what matters without leaving the page, picking the best model for each question.

Multi-model chat

Gemini, Claude, GPT, Qwen, and GLM — you pick the ideal model per conversation.

Summary and action items

Key points and tasks with owner, deadline, and priority — markable as done.

AI templates

Generate structured documents, always in the right format.

Repurpose and insights

Turn transcripts into posts, articles, and threads; see sentiment, recurring themes, and metrics.

10

Notebook

Ask everything at once

A multi-source workspace for deep analysis: bring several transcripts, PDFs, texts, and URLs into one context and question them together.

Combined sources

Transcripts, PDFs, pasted texts, and URLs in a single AI context.

Chat personas

Standard, Study, Critical, and Executive — each with specialized instructions.

Studio tools

The analysis and creation tools available over the combined sources.

Notes and search

Highlights and notes saved in the session, with a relevance search across everything you added.

Knowledge base

A searchable base of your own content (RAG), to query and cross-reference transcripts over time.

11

Integrations and TranscriBot

Take the transcript to where the team works

26 connectors to bring audio in and return the result to the right channel. Each connector states its connection type and availability.

MeetingsTranscriBot

Zoom, Teams, and Meet — the bot joins the meeting (with a valid calendar, link, and permissions), records the media, and triggers the configured flow.

Calendar and email

Google/Outlook Calendar link scheduled events; Gmail/Outlook transcribe attachments automatically.

Storage and docs

Dropbox, OneDrive, Google Drive, Notion, Confluence, OneNote, and Google Docs to import and publish.

Productivity and CRM

Slack, Trello, Asana, Monday, Jira, Planner, To Do and HubSpot, Salesforce, Zoho CRM to send summaries, tasks, and context.

Automation

Zapier connects to 8,000+ apps to build any workflow without code.

12

Export

The right format, the right destination

Formats

DOCX, PDF, TXT, and SRT — plus extra formats where supported.

Direct destinations

Send straight to Google Docs, Dropbox, or Gmail.

Subtitles and timestamps

SRT and VTT with provider timestamps, subtitle translation in compatible flows, and word-level timestamps.

13

Advanced settings

Fine control, when the case calls for it

Accuracy and noisy audio

High accuracy (Beam Search) and a dedicated profile for low-quality recordings.

Knowledge area

Domain and subarea to guide vocabulary and formatting (e.g. Law, Health, Engineering).

Custom dictionaries

Create and reuse dictionaries of terms and spellings by area, applied to transcripts per provider.

Content understanding

Entities, sentiment, chapters, and summary from the provider (AssemblyAI, Gladia, ElevenLabs), depending on the engine.

Human structure review

Optional checkpoint before final formatting, with a review modal in the interface.

14

Workspace and security

Who has access, where data lives, how it is protected

Team and roles

Invite colleagues and set access by role in the workspace.

Data residency

Processing profiles: Standard (any region), Europe (GDPR), or Brazil (LGPD).

Protection levels

Standard, Confidential, and Restricted — the last two, with reinforced controls, on the Enterprise plan.

Encryption

AES-256 at rest and TLS in transit; model training off by default.

Admin dashboardsenterprise

Compliance controls, fraud detection, observability, usage telemetry, and user management.

15

Dashboard and management

One place to keep every transcription moving

Dashboard and files

Track what is processing and what is ready; every transcript ready to search and export.

Notes and calendar

Keep key points in one place and see scheduled meetings and transcriptions.

Templates and settings

A library of ready formats and defaults for transcription, formatting, and interface — set once, adjustable per file.

16

API and developers

Transcribe programmatically

Beyond the interface, Transcriwise exposes an API to automate transcription and formatting from your own systems.

Public API

Send audio and retrieve transcripts and formatted documents by request, with dedicated documentation.

Keys and access

Per-account credentials, with scope and access control.

Usage, limits, and webhooks

Usage tracking in the dashboard and webhooks to deliver results to external destinations when processing completes.

17

Plans and billing

Pay for what you use

Plans and credits

Subscription plans, standalone credits, and special programs (e.g. education and accessibility).

Dual metering

Usage combines weighted transcription minutes and AI points for formatting and analysis.

Pro and Business tiers

Advanced features (Studio, Pro models, study material, teams) scale by plan.