> ## Documentation Index > Fetch the complete documentation index at: https://docs.intelligence-management-platform.com/llms.txt > Use this file to discover all available pages before exploring further. # Transcript settings > Workspace defaults for transcription sources, the meeting bot, recordings and retention. export const RightsCallout = ({right, flags}) => Who sees this: requires {right === "none" ? "no special rights — available to every signed-in user" : `the ${right} right (or super admin)`}. {flags && flags !== "none" ? ` Only visible when the server enables ${flags}.` : ""} ; Everything on the Transcripts page draws its defaults from here. Open it from **Transcripts → Settings**; the page is marked **Admins only** and other users see "Ask an admin" instead of the controls. Each setting saves on change, and any setting left alone shows where its value comes from — *"Using the environment default"* when a deployment variable supplies it, or *"Using the built-in default"* when nothing does. **Reset to default** returns a setting to that inherited value rather than to a blank. ## Sources The three ways a transcript can be created, each with a **Test** button that reports **Connected**, **Not configured** or **Unreachable**: * **File upload (Whisper)** — uploading an audio file * **Meeting bot (Recall)** — sending a bot to a meeting * **Record on this device** — recording the microphone in front of you A source that is not configured simply does not appear in the composer, so this section is where to look first when someone reports a missing option. ### Make transcripts searchable Transcripts need an **embedding model** before they can be found by Ask-about-this-transcript or by knowledge search. Without one, transcripts are still created, corrected and exported — they are just invisible to search. The section header warns when no model is set, which is the single most common reason "ask about this transcript" returns nothing. ## Meeting bot The name the bot joins under, and whether it announces in chat that the meeting is being recorded. **Recorders may set their own bot name and notice** decides whether those two are a workspace rule or a per-meeting choice. With it off, every bot uses the settings here even if a recorder tries to change them, and the composer tells them so. Announcing the recording is a legal requirement in many jurisdictions. Turning the notice off is a decision for whoever owns that risk in your organization, not a convenience setting. ## Defaults * **Default sharing for new transcripts** — the sharing a transcript starts with. It takes effect when the transcript is published; see [Sign off, and publish](/user-guide/transcripts/reviewing-and-saving#sign-off-and-publish). * **Default summary presets** — pre-selected in every composer, and still removable for a single recording. A preset whose prompt or agent has since been deleted is flagged here rather than failing quietly later. ## Recordings & storage * **Meeting video retention** — how long meeting video is kept, or **Keep forever**. * **Keep a local copy of meeting video** — when off, playback uses a temporary link from the meeting-bot provider instead of storage you pay for. * **Estimated storage cost per hour** — a rate you supply. IMP multiplies it by the hours stored this month and shows the estimate; it bills nothing and reads nothing from your provider. ## Recording limits An optional **monthly recording cap**, in minutes, or **No cap**. ## Knowledge base and agents Which agents can reach the transcripts knowledge base, and whether each can read, write, or write without reading. This answers "who can see what we record" in one place rather than agent by agent. An empty list is meaningful: **no agent uses this knowledge base yet, so transcripts are invisible to every agent** no matter how they are shared. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.