Audio to text
A candidate speech model would separate speech from silence, transcribe each turn, and preserve speaker, time, and words. Vocero's own model is not trained yet.
CareCompile speech program
From a meeting to reviewable minutes
Vocero proposes a workflow from audio to transcript and from transcript to minutes a person can verify. Every decision, task, or answer should retain a link to what was said.
In development for internal use. Our own model is not trained yet.
01 / VOCERO
The proposed workflow
The architecture separates speech recognition from meeting analysis so each stage can be evaluated on its own.
Text to structure
A candidate language model would extract entities, decisions, tasks, owners, and open issues from the transcript. This stage is also part of the plan.
Structure to minutes
The planned output organizes findings into readable minutes and retains references to transcript lines for review.
02 / VOCERO
Evidence before saving
The minutes assist the person responsible for the meeting; they do not make an automatic decision.
Source citations
Every point should identify the line or lines that support it. If the meeting does not contain an answer, the system should say so instead of inventing one.
Human review
A person confirms names, numbers, decisions, owners, and dates before saving or sending the minutes.
Visible limits
Names, numbers, or decisions absent from the transcript count as inventions and must be recorded in evaluation.
03 / VOCERO
What the demonstration shows
The interactive presentation uses fictional meetings with known correct output to explain the format.
A duplicate invoice charge
A synthetic customer-support script shows a duplicate charge, a priority case, a credit note, and confirmation by email.
Questions with sources
Scripted answers cite the relevant lines and identify missing information, such as the case number.
Format, not performance
Timings, waveforms, entities, and answers are programmed. The example is not a test of speech recognition or Vocero's own model.
04 / VOCERO
Current status
The application is used internally with a dictation engine that is not Vocero's own model. The program still needs to create data, train candidates, and measure results.
Today
The site and its synthetic demos run, and CareCompile internally uses a meeting and dictation application.
Pending
There is no trained Vocero model, no accuracy figure, and no completed sealed evaluation.
Explore the program
Related pages
Dictation for regional Spanish
Vocero investigates dictation for Salvadoran and Central American Spanish. Try the implemented command rules and explore the regional speech plan.
Explore the scopeA counterpart for Salvadoran voices and texts
CareCompile seeks an institutional counterpart in El Salvador to explore consented speech and text data, limited purpose, and independent evaluation.
Explore the scopeArchitecture and evaluation before claims
Review Vocero's proposed architecture, training plan, and Vocero-Bench design for measuring speech, dictation, and minutes by country. There are no results yet.
Explore the scopeEL SALVADOR / CARECOMPILE
Designate an institutional counterpart in El Salvador
We are looking for an institution willing to discuss regional meetings, voices, and texts. We explain purpose, consent, and governance before recording or receiving data.