What's new in 1.1.

1Chat, much smarter

Chat in 1.1 is more intelligent, faster and simpler. It understands the whole interview rather than a slice of it, tells the difference between a question and a request for clips, answers in plain words, and thinks like an editor sitting next to you. Every question, from "what's the story here?" to "suggest a running order", gets a better answer than before.

It reads the whole interview, however long it is. Long interviews (over an hour) used to get a list of clips and no answer. Now every question gets a real answer in plain words, with playable clips inside it where they belong. On a 3½-hour interview an answer arrives in about 20 seconds instead of one to three minutes, and there is no more "couldn't find moments" dead end.

It looks like a normal chat. One box, five suggested questions to start, nothing else on screen until you ask. The suggestions are about finding the story ("What's the story here?", "Where does this interview turn?", "Pull the strongest soundbites", "What are the major themes?"); once you've pulled clips, they turn toward your cut ("What am I missing?", "Suggest a running order").

Answers are short and specific. No essays, no headings. A few sentences from a sharp editor, then the clips that back it up.

Every clip in an answer is a card you can use. Play it in place, read its words, add it to Clips with one click, or pull all of them at once. "Build as Story" turns an answer into a Story Builder sequence.

It knows when not to give clips. Say "no clips" or "just tell me" and you get prose only. Ask something the footage doesn't cover and it says so, without bolting clips on. Ask something off topic and it answers plainly and stops.

Follow-up questions. Three suggested next questions appear under every answer; click one to ask it.

It remembers. Tell it what you're making ("a 90-second teaser", "a four-minute recruiting film") and every later answer keeps that in mind. Click ✕ on any clip card to pass on that moment; it won't come back in later answers, and you can restore it any time.

Little things that add up. Cards always have a real title. Hover a card to park the player on that moment before you press play. The progress line just says "Reading…". Speaker names appear in answers instead of "SPEAKER_04". Your active My Style profile shapes which moments it picks, not only how it talks. And it works well on every Mac tier, including 8 GB machines.

2Correct the transcript by double-clicking a word

Transcripts are good, not perfect. A name spelled wrong, a word the engine skipped, two speakers run into one paragraph, or a line credited to the wrong person used to mean living with it or re-transcribing. In 1.1 you fix it in place, in the Transcript tab, without touching the media.

Double-click any word to edit it. A small box opens on the word. Type the correction, press Enter, and the transcript, the paragraph and every clip that covers that moment show the new wording. Escape cancels.

Insert a word the engine missed. In the same box, Insert after opens an empty slot right after the word; type the missing word and press Enter.

Split a paragraph where it should break. Split here starts a new segment at that word. A small picker asks who the new part belongs to: keep the same speaker, pick any speaker already in the interview (shown by name), or New speaker. Rename a new speaker in the Speakers panel as usual.

Give a segment a different speaker. Speaker in the same box opens the same picker for the segment under the word, including on interviews where speaker identification has already run. A speaker you assign by hand stays: re-running speaker identification later leaves your assignments alone.

Undo. Every correction goes on its own undo stack. Cmd+Z steps back one correction at a time, separate from the label Undo button. Undoing everything returns the transcript to exactly what it was.

Media timing never changes. A correction changes words, never when they were said. Clips, selects, soundbites, story beats, exports and the player all keep their timecodes. Text copies that live on your clips and in the AI Analysis (a select's words, a soundbite's quote, a quote sheet's raw text) are refreshed for the range you edited; your own cleaned-up quote text is never touched.

It tells you when the analysis is behind. After a correction a small line above the transcript says the transcript was edited since the last analysis, with a Re-run button. Nothing re-runs on its own. Chat sees the corrected wording straight away.

3Collections work like a project

A collection is two or more interviews filed together: the founder and the investor, the three people in one campaign, the whole day of a shoot. In 1.1 a collection has the same page as a single interview, and every tab works across all of its interviews at once. The story you're looking for usually lives between interviews, and now the whole app looks there with you.

Create one from the Projects page. Every transcribed interview has a checkbox. Check two or more, click Create Collection, give it a name, and the interviews are filed together and the collection starts building. No more dragging projects into a folder first (dragging still works). Add an interview later and the collection picks it up.

Same tabs, same order. Transcript, Chat, Story Brief, Story Builder, Clips, Export. The old Dashboard is gone; everything it showed lives on the Story Brief tab, where it belongs.

Transcript. The same player as a project: video, or the live waveform for audio-only interviews, the audio-only switch, the same speeds. Switch between interviews with one click. The same color-label toolbar, and the same Speakers panel: colored rows you can rename and re-run speaker identification.

Chat across every interview. The same chat as a project, with the same five story-first questions, live answers, and clip cards that can come from any interview, each one naming its interview and speaker. Transcript on every card, Pull Clips and Build as Story, follow-up questions, "no clips" respected, and Story So Far: what you're making and the moments you passed on. Ask "pull the strongest soundbites" or "what are the major themes across these interviews" and the answer draws on all of them.

Story Brief across interviews. Arcs, tensions, strongest moments, the connections between interviews, speakers and coverage, each moment labelled with the interview it comes from, with Play, +, Transcript and Send to Story Builder. Rebuild it from the tab whenever an interview changes.

Story Builder. Drag clips in from Your Clips on the left, rename or delete stories in the Stories list, add with + Story, and give it a duration target ("a 3-minute story") that is measured against the result.

Clips. Every clip has a title, its first line and Show text; drag to reorder; a selected count; a scrub bar; color filters on one line.

Export. A timeline name field with the automatic "Collection – Selects 1" numbering, Documents (Word, PDF, Excel with a section per interview), AI Social Clips and AI Soundbites categories, Cuts + Markers, and the Edit-in choice remembered per collection. Collection exports default to Auto frame rate so 29.97 media stays on its own grid.

Header and settings. The same header and gear menu as a project: Collection details (rename, the interviews in it), Output Language, AI Model, Delete collection, and Assistant access for every interview in the collection at once.

Speaker names everywhere. Names found in an interview flow into the collection's transcript, chat answers, Story Brief and exports.

4AI assistant access: a Doza Assist connector for Claude (MCP)

Your projects can now be read by an AI assistant you already use, such as Claude. Your media files never leave your Mac; the transcript text and selects of the projects you switch on do go to that assistant, which is a cloud service.

Per project, off by default. On the project page, open the gear menu and choose AI assistant access. Only projects you switch on are visible to a connected assistant; everything else stays invisible.

Or per collection. On a collection page the same gear-menu item switches every interview in the collection on or off at once, and the assistant sees which collection each interview belongs to, so it can work across all of them for one story.

Add it to Claude Desktop with one click. The panel's Add to Claude Desktop button hands Claude Desktop the connector; Claude asks you to confirm, and Doza Assist appears in its connector list with the Doza Assist logo. No config files, no restart. Then ask Claude: "list my Doza projects". Claude Code, Cursor and VS Code connect with the command and config shown under Other assistants in the same panel. (ChatGPT cannot connect in this version.)

The panel says it plainly. One switch, Claude can read this project, and under it the facts: Claude runs in the cloud, so when the switch is on the transcript text and selects go to Anthropic's servers whenever Claude reads them, your media files stay on your Mac, and if the work is confidential you leave it off.

What a connected assistant can do:

• list the projects you've switched on,
• read a project's transcript with timecodes and speakers,
• search a transcript,
• see the selects in the Clip Library,
• add its own selects, clearly labeled "AI assistant" so you always know which are yours; they appear in the open project within seconds, with a toast and a View button,
• build a selects stringout you can open in your editor.

What it cannot do: delete or change your clips, see your media files, see projects you haven't switched on, or read your My Style profiles.

What leaves your Mac, and what doesn't. The connection itself runs on your Mac. When you switch a project on, the assistant can read that project's transcript text and selects, and whatever it reads is handled by that assistant's own service under its own terms: Claude is Anthropic's cloud service, not part of Doza Assist. Your video and audio files never leave the machine, and projects you haven't switched on are never visible. Access stays on until you turn it off. If a project must stay entirely private, leave it off.

Well behaved alongside other tools. The connector never interferes with servers that belong to other apps, says "starting up" while the app is still loading instead of "not running", and explains an empty project list (no projects yet, or none switched on). If two editions of Doza Assist are installed on one Mac, they recognise each other's connector as the same release instead of offering to "update" it back and forth.

Works in the trial. Assistant access is available in the trial on the two-minute trial transcripts, so you can try the workflow before buying.

5A cleaner interface

Small changes you'll notice on the first launch:

• Chat opens as a single box with five suggested questions, nothing else on screen until you ask.
• Tabs run in the recommended order of work, left to right: Transcript, AI Analysis, Chat, Story Brief, Story Builder, Clips, Export. Follow them across and you've made the film. They also lose the explainer sentence at the top.
• Project gear menu is ordered by scope and shows the current value of each setting; it gains Project details (name, client, interviewer, subject, speakers) and Relink media.
• Header shows a neutral "Style: Off" chip instead of a red warning; the appearance toggle is a single icon.
• AI setup lists the local model first and says clearly when a choice would send project text to a cloud service.
• First-time setup shows real download progress for both downloads, says what each download is for in plain words, counts the steps, and tells you the first start takes about a minute.
• Clips everywhere carry a title, a first line and Show text (details in section 7).
• Toasts say where clips went, with a View button, and no longer stack.
• Dark by default, and it stays put. Every page starts dark, applies the appearance you chose before it paints, and a new collection or batch page no longer opens in the other one.

6Exports named the way you'll look for them

• Every timeline arrives named for the project and what it is: "Project – Selects 1", "Project – Markers 1", "Project – Story: Title". The number counts up per project, so re-exports never pile onto one name. The file carries the same name.
• The Export tab lets you type a different timeline name before you send.
• Round-trip exports keep your original event.
• Every exported clip carries its transcript. In Final Cut, the clip's Notes field holds the title and speaker, then the exact words for that range, speaker by speaker, so you can read a select without playing it. Clips are tagged with one "Doza Assist" keyword for filtering.
• Story-arc exports open in Final Cut again: drop-frame camera footage exported from a Collection used to land with a timecode flag Final Cut rejects. Timecode now follows the timeline's frame rate.
• Story Builder sequences can be exported straight from the Export tab.
• Quote sheets match the press format comms teams send out: speaker lines, "On topic:" lead-ins, verbatim quotes and a Q&A index with time ranges. The previous branded layout is one setting away.
• Transcript corrections travel with the export: the note on a select shows the corrected words, on the original timecodes.

7Clips read like clips, everywhere

• Every clip has a short generated title, the first line of what's said under it, and Show text to read the whole passage: in the Clips tab, the Story Builder sidebar, the Story Brief, and every row of AI Analysis (soundbites, story beats and social clips).
• Trim a clip one word at a time. The Trim button on a clip opens a trim sheet: the words around the clip with in and out handles. Drag a handle, click a word, or nudge with the arrow keys, and the clip snaps to word boundaries. Audition the new edge in place.
• Clip cards show seconds, not a rounded minute.
• Clips start and end on the word. Timecodes keep sub-second precision, so a clip no longer opens a beat early on the tail of the previous sentence.
• Drag clips from the Story Builder sidebar straight into a story.
• Chat clip pulls say where the clips went ("Pulled 3 clips to the Clips tab") with a View button.
• Selects an AI assistant adds appear in the open page within seconds, clearly attributed, and the attribution survives every later save.

8Speakers

• Names come from the transcript. After speaker identification the app reads the interview for introductions and direct address and names the speakers it can, upgrading "Speaker 1" style labels to real names. Names are only ever taken from what's said; the app never invents a name or a description, and a name with a role reads like a lower third: James Casey, Superintendent of Schools. Stray tags such as "[name + role]" or fragments of the model's reasoning no longer appear as names.
• Naming runs on its own after speaker identification; the separate "Find speaker names" button is gone.
• Your assignments win. A speaker you set by hand on a segment (section 2) survives re-running speaker identification.
• Speaker names flow everywhere: Collections, the Story Brief, chat answers, exports.
• Speaker identification picks up where it left off after the app restarts, and a Run speaker detection button appears for projects that never ran it.

9Story Brief

• The Story Brief tab sits next to Chat and reads Arcs / Tensions / Coverage, shows when it was built and with which style profile, and asks before rebuilding.
• Each moment names its real speaker and opens a word-for-word transcript.
• One click adds a single moment to your clips.
• If the model's reply can't be read, the brief retries once and never replaces a good brief with an empty one. The Rebuild button returns to normal when the rebuild finishes.

10Transcription and analysis

• Auto-detect language listens to the first stretch of real speech and identifies the language before transcribing; English goes to the fast engine, whether you add one file or several at once.
• Very long recordings transcribe. A 36-hour file no longer runs out of memory partway; when the fast engine can't handle a file, the app says so.
• Guardrails before you wait: a warning when one project runs past 8 hours (4 hours for languages other than English), with a suggestion to split into shorter projects and use a Collection; and a warning when an FCPXML's source media adds up to more than twice its timeline, which usually means camera and recorder files were never synced. Both offer Continue anyway.
• AI Analysis is faster on long interviews. The model keeps its place from one chunk to the next instead of re-reading, and analysis and chat share one context size, so switching between them no longer throws away a warm model.
• A queued analysis says which analysis it is waiting for and how far along that one is.

11Settings and project details

• The project gear menu is reordered by scope and shows the current values.
• Project details lets you edit the name, client, interviewer, subject and speakers.
• Relink media when a source file has moved.
• New projects always start on the local model. If you switch a project to a cloud assistant, the app asks first and says what will leave the Mac.
• AI setup lists the local model first; the appearance toggle is a single icon; the header shows a neutral "Style: Off" instead of a red warning.

12Fixes

• Long interviews: when a chunk's analysis reaches the model's output limit, everything the model finished is kept and the cut is logged, instead of the tail being lost silently.
• Long interviews: chat stays grounded on a well-worked project. When the transcript plus a long conversation would not fit the model's window, that turn is answered from retrieval instead of losing the start of the transcript.
• AI assistant connector: a license activation no longer drops a connected assistant mid-call.
• Fewer repeated requests while you work: one chat warm-up per tab switch, one clip-titles request at a time, and one audio-track probe per source file per page (a probe reads the whole media file).
• In a collection, clip cards in chat take their title from what the speaker says, not from the interviewer's question, and a moment cited twice shows once.
• In a collection, a chat question no longer appears twice after you come back to the conversation.
• In a narrow or compressed window the player no longer sits on top of the text below it; it stacks above, with the picture capped so the controls stay in view.
• Retranscribing a project and then running AI Analysis no longer leaves the analysis tab empty (the cached analysis is restored and the derived files are rebuilt without re-running the model).
• MPEG-2 camera files (Sony XDCAM MXF, broadcast MPEG-TS) now report their frame rate correctly. The frame-rate picker and every export used to default to 23.976 for those files unless you changed it by hand.
• A brand-new folder no longer disappears when you drag the first project into it, and never nests inside the Unfiled area.
• "Send to Final Cut" opens the Final Cut you actually use when two copies are installed.
• Internal project ids never appear in analysis text, coverage gaps or story cards.
• Story Builder cards in a Collection open a real transcript, show seconds, carry their own title per beat, and keep each tension's own quote.
• Toasts no longer stack.
• The explainer sentence at the top of each tab is gone; tabs are cleaner.
• Story Brief arc chevrons point the right way.
• Collections default to Auto frame rate.