# Send Feedback to NexLev

> Report a bug, suggest an improvement, or request a feature straight from Claude, ChatGPT, or any MCP client. You or your AI assistant can file it. 5 per month, free on every plan.

Section: NexLev MCP  
Badge: New  
Canonical page: /docs/mcp/feedback

---

Send feedback to the NexLev team without leaving the chat. Report something broken, suggest how a tool could work better, or ask for a feature that doesn't exist yet.

Feedback can come from **you** or from **your AI assistant**. Ask the assistant to pass along what you said, or let it file a report itself when a NexLev tool returns wrong, confusing, or missing data, or when none of its tools can do what you asked.

## Prerequisites

- **Connected AI assistant**: Claude, ChatGPT, or any MCP-compatible client. See the [Quickstart](/docs/mcp/quickstart).
- **Signed in to your NexLev account**: anonymous connections cannot send feedback.
- **Any plan, including Free.**

> **Warning:** **Submit-only.** After you send feedback, you can't list, edit, or delete it
> from the assistant. Read the draft before it goes, especially since you get 5
> per month.

## Tool

### `nexlev_feedback_for_agents_humans`

Sends one piece of feedback to the NexLev team: a bug, an improvement, a feature request, or anything else.

**Best for:**

- "This tool gave me numbers that don't add up"
- "I wish this tool could also…"
- "Nothing in NexLev does X. Please add it"
- Your assistant reporting a tool that returned an error or empty results it shouldn't have

**Example prompts:**

```
Tell NexLev that similar-channel search ignored my Shorts filter. I asked for
Shorts-only channels like @channelhandle and got long-form channels back.
```

```
Write up the problem we just hit as a bug report: the tool, the exact
arguments you used, what we expected, and what came back. Show me the draft
before you send it.
```

```
File a feature request: I want frame extraction to accept a video's chapter
list instead of timestamps I have to type. Explain that I study one frame per
chapter to map how a video is structured, so it would save a step on every
video.
```

```
For the rest of this chat, if any NexLev tool gives you an error, empty
results, or data that contradicts another tool, tell me and offer to file a
bug report. Don't send anything without my OK.
```

```
I've mentioned three things today: the Shorts filter issue, wanting batch
metrics for more than 10 channels, and wanting results as a spreadsheet.
Combine them into one clear feedback note so it only uses one of my monthly
submissions.
```

**Parameters:**

| Parameter      | Required | Description                                                                                                              |
| -------------- | -------- | ------------------------------------------------------------------------------------------------------------------------ |
| `category`     | Yes      | `bug` (broken or wrong), `improvement` (an existing tool could be better), `feature_request` (something new), or `other` |
| `title`        | Yes      | A short summary, **3–150 characters**                                                                                    |
| `message`      | Yes      | The full note, **10–4,000 characters**. For a bug: which tool, what you asked, what you expected, and what happened      |
| `submitted_by` | No       | `human` if it is your words, `agent` if the assistant wrote it. Defaults to `agent`                                      |
| `related_tool` | No       | The NexLev tool it is about, for example `search_videos`                                                                 |

**Example call:**

```json
{
  "category": "bug",
  "title": "search_videos returns no outliers for a channel that has them",
  "message": "Asked search_videos for outlier videos from channel UC_EXAMPLE. Expected the channel's 3 known outlier videos, which the NexLev dashboard shows. Got an empty list.",
  "submitted_by": "agent",
  "related_tool": "search_videos"
}
```

## Write Feedback That Gets Acted On

Be specific. A report the team can reproduce is far more useful than a complaint.

| Weak               | Strong                                                                                                     |
| ------------------ | ---------------------------------------------------------------------------------------------------------- |
| "It doesn't work." | "`search_videos` returned no outliers for a channel that has them in the dashboard."                       |
| "Search is bad."   | "Similar-channel search ignored my Shorts filter and returned long-form channels."                         |
| "Add more stuff."  | "Let frame extraction take a chapter list, so I can grab one frame per chapter without typing timestamps." |

**For a bug, include:**

1. **The tool**: put it in `related_tool` too
2. **What you asked**: the prompt or the arguments
3. **What you expected**
4. **What actually happened**: the error text or the wrong result

**For a feature request**, say what you are trying to get done and why it matters, not just the feature. "I research 20 channels a week and compare them by hand" tells the team more than "add export".

> **Warning:** **Never include passwords, API keys, or other secrets** in feedback. Channel
> handles, video links, and tool arguments are fine.

## Allowance

**5 submissions per account per calendar month (UTC).**

- The cap is **per account, not per API key**: every connection on your account shares the same 5.
- At 5, nothing else is saved until the month resets, and the reply tells you the reset date.
- **Free on every signed-in plan**, including Free.
- **Does not count** against your weekly MCP tool-call allowance.
- **Never uses usage credits** and never bills.

| Plan                          | Feedback per month |
| ----------------------------- | ------------------ |
| Free                          | 5                  |
| Lite / Premium / Niche Finder | 5                  |
| Pro / N8N Pro                 | 5                  |

> **Tip:** **Make 5 go further.** Group related points into one submission, and ask your
> assistant to show you the draft first. Its own reports count toward the same
> 5, so tell it to check with you if you want to keep them for your own
> feedback.

## FAQ

### Can my AI assistant send feedback without me asking?

Yes. The tool is built for both people and agents. When a NexLev tool returns wrong, confusing, or missing data, or when nothing it has can do what you asked, your assistant can file a report on its own and marks it `submitted_by: agent`. If you'd rather approve every report, tell it to ask first.

### Can I see or edit feedback I already sent?

No. Feedback is submit-only. It cannot be listed, edited, or deleted from the assistant afterwards.

### Does sending feedback use my tool-call allowance or credits?

No. Feedback has its own limit of 5 per month and does not touch your weekly MCP tool-call allowance or your usage credits.

### What happens when I hit 5 in a month?

Further submissions are not saved until the next calendar month (UTC). The reply tells you the reset date.

### Is this the right place for account or billing problems?

For anything you need a direct answer on, like access, plan, or billing, email <contact@nexlev.io> or check [Troubleshooting](/docs/troubleshooting/overview). Use this tool for product feedback: bugs, improvements, and ideas.

## What's Next?

- **[Quickstart](/docs/mcp/quickstart)**: Connect NexLev to your AI assistant
- **[Video Frame Extraction](/docs/mcp/video-frame-extraction)**: Study editing, animation, and camera style frame by frame
- **[MCP Troubleshooting](/docs/troubleshooting/mcp)**: Fixes for common connection and quota issues