Bexar County Criminal Defense
The Evidence Orchestrator
It does more than transcribe. It reads the body-camera MP4 itself with Google Gemini and gives back an interpretation of the recording, the visual scene and the audio tone of the room, every speaker named, the line a document-only legal tool cannot cross.
Inside the app
A transcript is text on a page. This is a reading of the recording
Many tools give you a transcript. The Evidence Orchestrator does more than transcribe. It reads the body-camera MP4 itself with Google Gemini and returns an interpretation of the recording: the visual scene, meaning what the camera shows, and the audio tone of the room, meaning how it sounds, who is calm and who is raised. Every speaker is named as a character in the account: the officer, the complainant, the defendant, a bystander. Each moment carries the running time and the wall-clock time. And the whole thing comes back as a written account of what the footage actually shows, including the silences. That reading of the visual and audio environment, plus a character-named transcript, is exactly what the report delivers.
A document-only tool reads the police report, the affidavit, the typed witness statement, and it is genuinely good at that. A written report gives you the words. The Evidence Orchestrator gives you the scene on screen, the tone of the room, and the silences the words leave out, because it reads the footage itself rather than an account of it.
The line a document-only tool cannot cross
Here is the line, and it is the whole product. The Evidence Orchestrator does not read a summary of the footage. It opens the raw body-camera MP4 and reads the recording itself. It describes the visual scene the camera captured. It reads the audio tone of the room, the calm voice and the raised one. It names each speaker as a character so the transcript reads as people talking and not as anonymous lines of text. The MP4, the WAV, the JPEG: the actual things produced in discovery, read as the primary evidence they are.
That is the difference between a tool that can summarize a discovery letter and a tool that can show you the full scene: the man sitting on the curb for six minutes, the room quiet, no voice raised, all of it on the record beside the words. One gives you text on a page. The other gives you a reading of the recording. A defense built on the second one is a defense built on the record.
What the AI gives back
Point it at a matter on the Case Board and it returns an indexed, court-ready record: a synchronized timeline of every clip, a character-named transcript of every spoken word with the speaker and the time it was said, a reading of what the camera saw and how the room sounded, and a plain-language account of each photograph. Silence is recorded too, the long minutes where the camera films an idling engine and no one is read their rights, because what did not happen on the recording is as much a part of the defense as what did.
The reading is done by OpenAI’s ChatGPT, Anthropic’s Claude, and Google’s Gemini, with Google Gemini reading the video and audio, and the result is built to be cited. Jump from a finding straight to the second of footage it came from, the way you would cite a page in a brief. Nothing is invented and nothing is summarized away. Where the camera shows an interaction, the record says interaction. Where the room is quiet, the record says so, and names the quiet.
Built so it holds up at the bench
Picture the worn stone hallway outside the courtroom, the judge’s name on the door, your client waiting on the bench. The record the Evidence Orchestrator hands you is built for exactly that room: timestamps that match the body-camera clock, captions on the silent clips, a character-named transcript, and an index a judge or a prosecutor can open and check against the original media. It is a reading of the evidence the court can verify against the evidence itself.
This is what carries straight into the Article 39.14 discovery demand and the motion to suppress. The Issue Spotter reads what the Evidence Orchestrator found and builds the motions from it, so the body-camera evidence review and the filing are one continuous line of work, not two. It is also the record the Issue Spotter reads to answer the question that sits under the whole case: does the footage support the offense on the charging instrument, or does the recording fit a lesser charge? That charge-level read is built on this reading of the recording, not on a written summary of it.
One matter, one system
The Evidence Orchestrator does not stand alone. The case begins on the Case Board, where the client, the cause number, the court, and the appointed-counsel fee voucher that pays you already live, and the evidence review pulls straight from that matter. It is the second act of a three-act case: the Case Board where the matter begins, the Evidence Orchestrator that reads the recording, and the Issue Spotter that turns the record into filings ready for eFileTexas.
For managed assigned counsel and the indigent defense bar, this is the point. The appointed defender finally has the same reading of the recording that a fully resourced office would have, the visual scene and the audio tone and a transcript with every speaker named, in the one place where the rest of the matter already runs. That is what the Sixth Amendment promise reads like in 2026.
Private investigators run the same lane. Surveillance video, interview audio, and scene photographs are read the same way as body-camera footage, with the same reading of scene and tone and the same named-character transcript, and the confidentiality terms cover investigative work product the way they cover an attorney’s.
Four ways to read the clock
Every moment on the tape, fixed to time a jury can follow
A timestamp a juror cannot place is a timestamp that persuades no one. Each moment carries the tape time, the elapsed time, the wall-clock time, and the countdown to what happens next. Durations read three ways, too: as D:HH:MM:SS, in plain language, and against the wall clock.
Silence is a fact
The quiet on the tape is part of the record
A silent stretch of body-camera footage is evidence too. A clip with no speech is captioned for what it is, an engine idling, a door, ninety seconds of nobody reading anyone their rights, so the gap between what was said and what was done becomes something the record shows rather than something an attorney has to remember.
By hand, or in one run
Drag the divider
By hand
Someone watches every minute of every clip, in real time, and types what they hear. A day of body-camera footage is a day of watching, done at the end of a long docket, from memory.
In one controlled run
The same models, the same prompts, the same date read the whole set at once and hand back an indexed, jury-readable record. One run, so the dataset never looks stitched together.
Learn
Learn The Evidence Orchestrator
Short, plain reads, step-by-step tutorials, and a self-check on this pillar.
The whole machine
Ten logins down to two
One subscription stands in for the stack a defense firm usually stitches together, so the practice runs on two things it already trusts: Google Workspace, where the files and matters live, and The IMC Machine, where the matter runs. LawPay and QuickBooks, the money rail, stay where they are, fed by the Machine. Casework, evidence, motions, vouchers, marketing, and events, in one place.
San Antonio and Bexar County. The Machine is built for here first. For another city, county, or region, the build-out is custom and carries its own setup fee.
