# Edit Plan

> Plan an edit from a recording's transcript. Tighten it, find short shareable clips or mark chapters, as an edit decision list that Apply EDL renders.

Source: https://nodaro.ai/docs/nodes/video/edit-plan

The **Edit Plan** node reads the timed transcript of a recording and writes an edit decision list (EDL): a JSON plan that the [Apply EDL](https://nodaro.ai/docs/nodes/video/apply-edl) node renders into a finished cut. Edit Plan reads the transcript and never the pixels, so it plans quickly and cheaply. It has three modes: **Tighten** cleans up the whole recording, **Clips** finds short shareable moments, and **Chapters** marks chapters with titles.

- Found in: Video › Cut & Assemble
- Output: data
- API type: `edit-plan`

Edit Plan runs on Nodaro Cloud. On a self-hosted install, [connect the install to Nodaro Cloud](https://nodaro.ai/docs/self-hosting/cloud-connect), under **Integrations › nodaro.ai** or with an API key, and the node relays each run to the cloud. Without a connection, the node saves on the canvas but cannot run.

## When to use it
- Tighten a long podcast or interview recording before you render it.
- Find the strongest short moments of an episode and render each one as its own clip.
- Generate chapter markers with titles for a long upload.

## Quick start
### Add the node

Press Tab on the canvas and choose **Video › Cut & Assemble › Edit Plan**.

### Connect the transcript

Wire the JSON output of a [Transcribe](https://nodaro.ai/docs/nodes/audio/transcribe) node into the **Transcript** input. The transcript needs word-level timing.

### Connect the media

Wire the recording into the **Sources** input. You can connect 1 to 6 sources, for example several cameras and a separate microphone. In the settings panel, give each source its role.

### Choose the mode

Choose a **Mode** and a **Tier**. For **Tighten**, also wire a [Silence Detect](https://nodaro.ai/docs/nodes/audio/silence-detect) node into **Silence**, so the plan can cut dead air precisely.

### Run it and render the plan

Click **Run** on the node, then wire its **Edit plan (EDL)** output into an [Apply EDL](https://nodaro.ai/docs/nodes/video/apply-edl) node to render the cut.

Workflow: Edit Plan finds five short clips in a transcribed episode, and Apply EDL renders each clip as its own video.

- Upload Video → Transcribe (audio)
- Upload Video → Silence Detect (audio)
- Transcribe → Edit Plan (transcript)
- Silence Detect → Edit Plan (silence)
- Upload Video → Edit Plan (sources)
- Edit Plan → Apply EDL (EDL)

## Inputs
| Input | Accepts | What it does |
| --- | --- | --- |
| **Transcript** | JSON transcripts, such as the output of Transcribe | The timed, word-level transcript of the recording. Required. |
| **Silence** | The JSON output of Silence Detect | The silent ranges of the recording. Optional. They help **Tighten** cut dead air precisely. |
| **Sources** | Video and audio nodes | The media the edit draws from, 1 to 6 sources. Required. Each source becomes a source of the EDL. |

The output, **Edit plan (EDL)**, depends on the mode:

| Mode | Output |
| --- | --- |
| **Tighten** | One EDL: a cleaned-up version of the whole recording. |
| **Clips** | A list of EDLs, one per clip. The output fans out, so an Apply EDL node connected to it renders every clip, in parallel. |
| **Chapters** | A list of chapters, each with a start time in milliseconds and a title. |

## Settings
| Setting | What it does |
| --- | --- |
| **Mode** | **Tighten (clean up the recording)**, the default, removes silence, filler words and false starts. **Clips (find short clips)** finds the strongest short, shareable moments. **Chapters (mark chapters)** marks chapter starts with titles. |
| **Tier** | **Economy (fastest)**, **Standard** (the default) or **Premium (highest fidelity)**. The tier sets both the quality of the plan and its price. |
| **Instructions** | A short steer for the plan, for example `keep the intro tight, drop the sponsor read`. |
| **Style guide** | Longer editorial guidance that the plan should follow. |
| **Number of clips** | **Clips** mode only. How many clips to find, from 1 to 50. Leave it empty to let the planner decide. |
| **Target clip length (seconds)** | **Clips** mode only. The clip length to aim for, from 5 to 180 seconds. The planner aims near it, but it is not a hard limit. Leave it empty to let the planner decide. |
| **Target aspect** | **Any** (the default), `16:9`, `9:16`, `1:1` or `4:5`. The delivery shape, which informs how clips are framed. |
| **Sources** | The connected media, in order. Drag to reorder them, and give each source a role: **Auto** (the default), **Master audio**, **Camera**, **Wide** or **Screen**. |
| **Pre & post text** | Text that is always added before and after the instructions. It is hidden from people who use your workflow as an app. See [Prompt pre and post text](https://nodaro.ai/docs/concepts/prompt-pre-post-text). |

A source with the **Master audio** role gives its sound to every segment of the edit, so switching between cameras never changes the sound. See [Apply EDL](https://nodaro.ai/docs/nodes/video/apply-edl) for how the roles are rendered.

## Credits
Edit Plan is priced by the length of the recording and the tier. The length is rounded up to a bucket of 15, 30, 60, 90, 120 or 180 minutes.

`credits = rate per minute × bucket minutes`, plus a flat amount in **Clips** mode.

| Tier | Credits per minute | Flat amount in Clips mode |
| --- | --- | --- |
| **Economy** | 2 | 10 |
| **Standard** | 4 | 20 |
| **Premium** | 8 | 40 |

| Example | Bucket | Calculation | Credits |
| --- | --- | --- | --- |
| Tighten, Standard, 45-minute episode | 60 minutes | 4 × 60 | 240 |
| Chapters, Economy, 20-minute episode | 30 minutes | 2 × 30 | 60 |
| Clips, Premium, 90-minute episode | 90 minutes | 8 × 90 + 40 | 760 |

Rendering the plan with Apply EDL is priced separately, per minute of output.

### The estimate before you run

The cost on the node, on the **Run** button and in the run confirmation is an estimate. It uses the recorded length of the master source: the source with the **Master audio** role, or else the first source in the node's order.

- **Uploaded audio and video, and generated video,** carry their length, so the estimate lands on the right bucket.
- **A YouTube link extracted with [Reference Audio](https://nodaro.ai/docs/nodes/audio/reference-audio)** records the length of the extracted file, so the estimate is right too.
- **A source with no recorded length,** such as a direct audio link, is estimated at the largest bucket, 180 minutes. Extracting a YouTube source again records its length.

The estimate never takes the length from the wired transcript. That transcript comes from the previous run, and after you swap in a longer episode it would under-quote the new one. When the length is unknown, the estimate deliberately quotes high, so that a run is refused up front instead of failing halfway after earlier nodes were charged. Whatever the estimate shows, the charge is checked against the real length of the recording. The server measures the master source before it reserves credits, and a run whose master cannot be measured is refused and refunded.

## Tips
- **Wire the full chain.** Connect Transcribe to **Transcript**, add Silence Detect to **Silence** for **Tighten**, and connect **Edit plan (EDL)** to Apply EDL.
- **Render many clips from one plan.** In **Clips** mode, one Edit Plan node feeds many renders through a single Apply EDL node.
- **Draft on Economy.** Use **Economy** for a fast first pass, and **Premium** for the final plan.
- **Name the master audio.** When you record with several cameras and one good microphone, give the microphone source the **Master audio** role.

## From the API
The MCP `plan_edit` tool and `POST /v1/edit-plan` run the same planner with 1 to 6 media sources. The API also accepts a `platform` hint, which informs the pacing and the length of the plan. The finished plan is in the job result, ready to pass to `apply_edl`. See the [MCP tools](https://nodaro.ai/docs/mcp/tools) and [Run a single node](https://nodaro.ai/docs/developers/api/nodes).

## Frequently asked questions

### Does Edit Plan render a video?

No. Edit Plan writes a plan, an edit decision list. Wire its output into an Apply EDL node, which renders the finished cut.

### How much does Edit Plan cost?

The price is the recording's length, rounded up to 15, 30, 60, 90, 120 or 180 minutes, times a per-minute rate of 2, 4 or 8 credits for the Economy, Standard or Premium tier. Clips mode adds a flat 10, 20 or 40 credits.

### Can Edit Plan find several short clips in one run?

Yes. In Clips mode it returns one plan per clip, and the output fans out, so a single Apply EDL node connected to it renders every clip in parallel.

### Does Edit Plan work on a self-hosted install?

Yes, when the install is connected to Nodaro Cloud. The node relays each run to the cloud. Without a connection, you can place and save the node, but it cannot run.

### Why does the estimate show a 180-minute price?

The estimate uses the recorded length of the main source. When a source has no recorded length, such as a direct audio link, the estimate assumes the largest bucket. You are charged for the real length, which the server measures before it reserves credits.