# Run a cost-aware transcript pipeline

> Batch transcript pulls are free for imported videos; a credit check keeps the paid steps intentional.

- For: Data team
- Category: Agent & Workflow Automation
- MCP tools: library_import_youtube_video, get_youtube_transcript, get_credit_status
- HTML version: https://veridive.com/mcp/use-cases/free-tier-transcript-pipeline

## The problem

Data teams keep rediscovering the same need: a corpus of video transcripts — for analysis, for training material, for the internal search index, for whatever the quarter's project is. And they keep meeting the same cost anxiety: transcript acquisition at volume feels like a metered faucet nobody wants to be the one who left running.

The veridive cost structure has a shape worth designing around: **imports are the paid step; transcript pulls from imported videos are free.** That asymmetry is a pipeline design — pay deliberately once at the import gate, then read freely forever. Add a meter check at the gate and the faucet anxiety becomes a budget line.

## How veridive MCP helps

**library_import_youtube_video** is the paid gate: the corpus enters the library deliberately, batch by batch, tagged for the project. **get_youtube_transcript** is the free read: timestamped transcripts pulled per video, batchable across the whole imported set, re-pullable whenever the downstream format changes — at zero marginal cost.

**get_credit_status** keeps the gate intentional: check the meter before each import batch, size the batch to the budget, and the pipeline's spend is a decision made at one point instead of a surprise distributed across a thousand calls. The output: a tagged transcript corpus, acquired at a known cost, readable for free from then on.

## How to run it

1. **Connect veridive MCP** — One-time setup: add the veridive MCP server to Claude, ChatGPT, Cursor or any MCP client and generate your API key. Takes about two minutes.
2. **Meter, then import the batch** — The one paid step, sized to the budget and tagged to the project.

   ```
   Check my veridive credits first. If there's room for ~<N> imports, import these videos tagged '<project>': <video URLs> — use veridive's library_import_youtube_video.
   ```
3. **Pull the transcripts free** — The whole imported batch, timestamped, at zero marginal cost.

   ```
   Pull the full timestamped transcripts for everything tagged '<project>' — use get_youtube_transcript.
   ```
4. **Re-read whenever the format changes** — New downstream shape? Re-pull free — the import was the purchase; the reads are forever.

   ```
   Re-pull the '<project>' transcripts, this time formatted as <new format — e.g. plain text without timestamps>.
   ```

## Problems it solves

- Transcript acquisition at volume feels like an unmetered faucet — so projects stall on cost anxiety.
- Spend distributed across thousands of calls is unbudgetable; nobody owns the number.
- Re-processing needs (new format, new chunking) get priced as re-acquisition when they shouldn't be.
- Corpus projects leave no durable asset — the next project re-fetches everything.

## Why it's valuable

- The pay-once/read-free asymmetry becomes the pipeline's architecture.
- Spend concentrates at one deliberate gate, sized by the meter check.
- Re-pulls are free: downstream format changes stop costing acquisition money.
- The tagged corpus persists — the next project starts from the library, not from zero.

## Example prompts

```
Check credits; if room for <N> imports, import these tagged '<project>': <URLs>
```

```
Pull timestamped transcripts for everything tagged '<project>'.
```

```
Re-pull '<project>' transcripts as <format>.
```

## FAQ

**Is the transcript pull really free?**

For videos already in your library, yes — get_youtube_transcript is a free call. The import is the metered step, which is exactly why the pipeline gates there.

**What about videos we don't want to keep?**

The library is the mechanism, not a commitment — but for corpus work you generally want the imports to persist: that's what makes re-reads free and the asset durable.

**Does this compose with the caption workflow?**

Directly — the same imported corpus serves the SRT/VTT pipeline (see the captions use case) and any run_youtube_tool processing. Import once, feed every pipeline.

## Related use cases

- [Ship captions and subtitles without the grind](https://veridive.com/mcp/use-cases/caption-accessibility-workflow.md) — Pull SRT/VTT files and clean transcripts in one step — accessibility and search indexing in the same pass.
- [Give your research agent a spending limit](https://veridive.com/mcp/use-cases/budget-guarded-agent.md) — The agent checks remaining credits before each deep run and degrades to lighter tools when the budget thins.
- [Know your burn across research sprints](https://veridive.com/mcp/use-cases/know-your-burn.md) — Query balance and daily quota mid-sprint — plan heavy research days around what's actually left.

---

Part of the veridive MCP use-case library: https://veridive.com/mcp/use-cases.md
Server docs for agents: https://veridive.com/mcp.md