Three timed blocks — subject matter, tech, logistics — that turn a cold assignment into one you walk into calm. With medical, legal, and conference variants.
Ask working interpreters what separates a smooth assignment from a rough one and few say talent; most say what happened before it started. Preparation is the invisible half of the job — and because it's invisible, it's the half that gets squeezed when the booking lands at 5 p.m. for a 9 a.m. start. Thirty disciplined minutes, split into three timed blocks, covers what actually fails on assignments: unknown terminology, unverified audio, and no plan B. Set a timer for each block and stop when it rings — the constraint is what makes the routine survivable enough to run every time.
The goal of this block is not expertise — it's removing the words that would otherwise stop you mid-sentence.
Agenda, case summary, slide deck, patient context — whatever the client provided, and only that. You're mapping the terrain, not studying it: what is this meeting for, who talks, what's decided.
Every assignment turns on a small set of load-bearing vocabulary: names of people and organizations, technical terms, the acronyms everyone in the room already knows. Write them down — twenty is usually the honest number.
For each term, commit to a rendering in your other language — once, in writing, while calm. The booth is the wrong place to weigh two options for "stent placement" or a three-way ministry name.
If no materials came, this block becomes ten minutes of informed guessing: the organization's website, the speaker's last talk, the standard vocabulary of the setting. Thin prep beats none by a wide margin.
Audio is the assignment. Everything in this block exists because "can you hear me?" is not a question you want to be asking at minute one of a deposition.
Remote: join the actual platform with the actual headset and confirm you can hear and be heard. On-site: know what feed you're getting and from where. If you run live support, test the same path — for meeting audio that means system or tab audio, not the microphone hearing your speakers.
Language pair set, auto-detect on if speakers may switch, translation view arranged where a glance reaches it. Confirm the working mode — simultaneous, consecutive, or a mix — so your tools and your note-pad match the job.
Move the twenty terms from block one into Unicaption's custom dictionary so names and terminology render correctly from first mention. This is the highest-yield four minutes of the half hour: it converts your prep into machine accuracy.
On-site: address, room, arrival time, contact's phone number, badge or check-in requirements. Remote: the join link opened and bookmarked, dial-in as backup, the platform's rejoin behavior understood. Confirm session length and whether you get relief.
How you'll ask for a repetition or clarification, first-person convention, who to signal when audio degrades. Thirty seconds agreeing this with the client beats improvising it mid-testimony.
One line each for the three classic failures: audio dies (dial-in number ready), tech dies (pen and pad positioned, dictionary terms also on paper), terminology surprises you (the clarification phrasing you'll use). Fallbacks you've written down are fallbacks you'll actually use.
The three blocks hold across settings; where each block's minutes go shifts with the stakes of the setting:
| Block | Medical | Legal / court | Conference |
|---|---|---|---|
| 0–10 · Subject | Diagnoses, medications, dosages; the patient's context if shared | Case names, charges, statute and document names; exact spellings | Speaker bios, talk titles, company and product names |
| 10–20 · Tech | Telehealth platform check; built-in medical terminology on | Court audio feed; built-in legal terminology; consecutive protocol | Booth feed or RSI platform check; slide deck loaded beside you |
| 20–30 · Logistics | Session length, clinician contact, consent norms of the clinic | The record: who owns it, your correction protocol, oath wording | Relief rotation, handover signals, Q&A plan |
Court work adds constraints of its own — the record, the oath, strict device rules — covered in our court interpreting guide. Booth and RSI assignments layer team coordination on top; see how AI fits the conference booth.
Unicaption is designed to be the tech block's five-minute setup and the assignment's quiet second pair of eyes:
The free tier gives you 30 minutes every week with no credit card — fittingly, exactly the length of one full prep run. Load a real glossary, test your real audio path, and see what the transcript gives you.
The point of the timer is not efficiency; it's repeatability. An open-ended prep session expands into the evening on the assignments that frighten you and collapses to nothing on the ones that don't — and the ones that don't are where surprises live. A routine that costs exactly thirty minutes gets run every time, and every-time prep compounds: your glossaries accumulate, your fallback lines stop needing rewriting, and the assignment starts from calm instead of adrenaline. That last part is not a soft benefit — walking in prepared is one of the few levers interpreters control against the profession's cognitive load and burnout problem. The work is hard enough when only the work is hard.
Split the half hour into three timed blocks. Minutes 0–10: skim the materials, pull the twenty load-bearing terms, and decide your rendering for each in writing. Minutes 10–20: verify the audio path end to end with the real platform and headset, set your language pair, and load the glossary into a custom dictionary — Unicaption's takes it directly. Minutes 20–30: close the logistics, agree clarification protocols with the client, and write one fallback line each for audio, tech, and terminology failures.
The names and terms the assignment actually turns on — usually about twenty: people and organizations, technical or clinical vocabulary, acronyms the room already knows, and place or document names spelled the way the file spells them. For each, commit to one rendering in advance. Loaded into Unicaption's custom dictionary, the same list also makes the live transcript render those terms correctly from first mention.
Test the real path, not a stand-in: join the actual platform with the actual headset, confirm both directions, and run about sixty seconds of live audio through any transcription or support tools while watching the output. For meeting audio, the source should be system or tab audio — captured digitally — rather than a microphone listening to speakers. If the transcript is clean in the test, it will hold in the assignment.
Three, written down during prep: audio fails (a dial-in number or secondary feed ready), tech fails (pen and pad positioned, key glossary terms also on paper), and terminology surprises you (the exact clarification phrasing you'll use, agreed with the client in advance). One line each is enough — the point is deciding while calm so the mid-assignment version of you only has to execute.
Glossary loaded, audio path tested, transcript running — 30 minutes to calm. First 30 minutes of Unicaption free, no credit card.
Start Free Trial →