# Find the spoken moment before editing a video > Arcmira's n8n workflow returns up to three timestamped transcript passages. Editors still need to check the source, speaker, cut and reuse rights. By BIG CHANGE Editorial Published: 2026-10-07T15:16:13.493Z Updated: 2026-10-07T15:55:45.557Z Canonical: https://bigchange.ai/blog/find-spoken-moment-before-video-editing ![Conceptual charcoal illustration of an editor wearing headphones and looking at an anonymous speaker in source footage on a desktop monitor.](https://bigchange.ai/api/media/file/arcmira-spoken-source-review-hero-v1.png) AI-generated conceptual editorial illustration by BIG CHANGE. [Arcmira](https://arcmira.com/) has published a small n8n workflow for finding spoken passages in indexed YouTube videos. An editor supplies one topic and a UTC date window, then receives up to three timestamped passages with links back to their sources. The links lead to recordings the editor can review before making a cut. This guide follows the public workflow and documentation checked on October 7, 2026. BIG CHANGE has not imported or run it. ## The big change - **What changed:** Arcmira's public n8n example turns a topic and date window into a shortlist of up to three source-linked transcript passages. - **How it works:** An inactive, manually started workflow sends one search request and preserves both the passages and the search coverage metadata in one JSON item. - **Why it matters:** Editors can navigate to candidate moments without treating a transcript match as a verified quote, speaker, usable cut or licensed clip. ## 1. Set up one bounded search Download the [workflow JSON from Arcmira's public repository](https://github.com/arcmira/integrations/tree/master/examples/n8n-source-shortlist). In the n8n editor, use the upper-right three-dot menu and **Import from File**, as [n8n's import guide](https://docs.n8n.io/build/manage-workflows/export-and-import/) describes. Arcmira says the workflow is inactive and starts manually. Its recorded run used n8n 2.41.3 in an isolated environment; browser-editor import and n8n Cloud have not been verified in that record. Create an **HTTP Bearer Auth** credential with your own Arcmira API key. Enter the key as the token, without adding \`Bearer \`, and select that credential in **Search Arcmira**. Keep the key in n8n's credential store, away from workflow fields and exports. In **Search input**, enter one `topic` and the `after` and `before` dates. The example uses `creativity`, from August 1 through the start of September 1, 2026. The dates are UTC: `after` is inclusive and `before` is exclusive. Choose a window your Arcmira plan can access. Arcmira's [usage guide](https://arcmira.com/docs/usage-and-billing) says its Free plan withholds the newest 30 days of media. ## 2. Run once and read the coverage notes Execute the workflow manually and inspect **Build source shortlist**. The example makes one `GET /v1/search` request with a fixed limit of three. It returns one JSON item with up to three `candidates`. Each candidate retains the video ID, title, channel, publication date, passage text and start time, plus watch and original YouTube links. Available transcript source and speaker labels travel with the result. Read `requested_window`, `applied_window`, `partial`, `search_index` and `note` alongside the passages. They show what the service actually searched and whether it reports incomplete coverage. An empty list does not establish that nobody discussed the topic: the search covers Arcmira's indexed material, not every YouTube video. The example has no pagination or automatic retry. A refused request stops; resolve the error before deciding whether to run a new query. Arcmira's [verification record](https://raw.githubusercontent.com/arcmira/integrations/master/examples/n8n-source-shortlist/verification.json) reports one successful run on October 7, 2026, with n8n 2.41.3 and a Free account. It returned three passages, marked `partial` false, and reported a `catching_up` search index. Arcmira measured no credit or on-demand change around that run. This is the repository publisher's record of one execution. It does not establish that a later search will return the same passages or complete coverage, and BIG CHANGE has not repeated it. ## 3. Open the source before choosing a cut Use a candidate's watch link and timestamp to find the moment in the original recording. Play the surrounding context, check the transcript against what is audible, identify the speaker, and decide whether the passage says what the shortlist suggests. Arcmira explicitly leaves cut start and end values null. The timestamp points you toward the source; it does not select frame-accurate in and out points. Speaker labels may be missing or need verification. Before using any excerpt, establish the rights needed for that video and audio. The workflow's open-source license covers its code and instructions, not third-party recordings or transcript content. The public example does not animate footage, select edit boundaries, or render through Canvas or ffmpeg. Those are separate downstream production decisions after source and rights checks. ## Check usage before expanding the search The current [Arcmira API page](https://arcmira.com/transcript-api) and [plans and usage guide](https://arcmira.com/docs/usage-and-billing) agree that Free includes 1,000 credits per month. Both say the first five search passages in a request are free; additional search results cost 4 credits each. The usage guide also says Free excludes media from the newest 30 days and gates later list pages. Arcmira's verification record reports no credit change for its one three-result run, but that is a publisher-reported measurement, not an independent BIG CHANGE test or a general usage guarantee. Check the live account balance and current plan terms before making other billable searches or ordering Premium transcription. Premium transcription is outside this n8n example. ## Sources & further reading - [Arcmira's n8n source-shortlist README](https://raw.githubusercontent.com/arcmira/integrations/master/examples/n8n-source-shortlist/README.md) documents the import, credential, inputs, request, output and limits. The instructions and claims come from the workflow publisher. - [Arcmira's verification record](https://raw.githubusercontent.com/arcmira/integrations/master/examples/n8n-source-shortlist/verification.json) describes one publisher-run n8n 2.41.3 execution and its coverage state; it is not an independent test. - [Arcmira's plans and usage guide](https://arcmira.com/docs/usage-and-billing) and [transcript API page](https://arcmira.com/transcript-api) describe the current monthly credit allowance and free first-five search-passage rule; the API page specifies the 4-credit cost for each additional search result. They also document the separate endpoint and its limitations. - [n8n's workflow import guide](https://docs.n8n.io/build/manage-workflows/export-and-import/) describes **Import from File** in the editor menu. BIG CHANGE did not verify this particular workflow's import in a browser editor. ## Sources - [Arcmira n8n source shortlist README](https://raw.githubusercontent.com/arcmira/integrations/master/examples/n8n-source-shortlist/README.md) — Documents workflow install, inputs, one fixed-limit search request, output fields, limits, current pricing statement and the n8n/browser verification boundary. - [Arcmira n8n source shortlist verification record](https://raw.githubusercontent.com/arcmira/integrations/master/examples/n8n-source-shortlist/verification.json) — Publisher's record of one live n8n 2.41.3 run; states three passages, no partial result, zero measured credit delta, index catching_up, and untested browser import/Cloud. Not independent verification. - [Arcmira YouTube Transcript API](https://arcmira.com/transcript-api) — Current API interface, timestamped search passages, first five passages per request free, 4 credits per additional search result, monthly Free allowance and limitations. - [Arcmira Transcripts guide](https://arcmira.com/docs/transcripts) — Transcript fields, caption/Premium distinction, timestamps, speaker data and request states. Premium is not used in this example. - [Arcmira plans and usage](https://arcmira.com/docs/usage-and-billing) — Live plan table confirms 1,000 monthly Free credits; documents first five search passages free, 4 credits for each additional search result, Free freshness and pagination gates, and balance inspection. - [n8n export and import workflows](https://docs.n8n.io/build/manage-workflows/export-and-import/) — Official generic Import from File path in editor menu; not a verification of importing the Arcmira JSON. - [Arcmira homepage](https://arcmira.com/) — Company homepage; linked at first mention in the article.