> ## Documentation Index
> Fetch the complete documentation index at: https://docs.hireflix.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Create a position via chat

> Fill in one template and let your AI tool build a complete Hireflix position, including intro, outro, and question videos or images.

With the [Hireflix MCP](https://docs.hireflix.com/ai/mcp.md) or the [Hireflix skill](https://docs.hireflix.com/ai/skill.md) set up, you can build a new position just by chatting with your AI tool — no GraphQL required. It uses the `createPosition` mutation (plus the matching upload steps for any videos or images you attach) to build the position exactly as you describe it.

Copy the [template below](#the-position-template), fill it in, attach your files, and send it in one message. Your AI tool creates the position straight away.

> [!NOTE]
>
> Some AI tools reply with this template when you simply ask "I want to create a new Hireflix position", but that depends on the tool and isn't reliable. Copying the template from this page always works.

## The position template

> [tab] Template
>
> ```text theme={null}
> POSITION
> Name:            (required, max 50 characters)
> Tags:            (optional, comma-separated, max 5, each max 25 characters)
> Intro video:     (optional: file name, or "none")
>
> QUESTION 1
> Title:           (required, max 40 characters)
> Description:     (optional, max 1200 characters)
> Attachment:      (optional: "video: <file>" OR "image: <file>", or "none")
>
> QUESTION 2
> Title:
> Description:
> Attachment:
>
> (add more questions the same way, up to 50)
>
> OUTRO (optional, delete this block to skip it; Title and Text are required if you keep it)
> Title:           (max 100 characters)
> Text:            (max 500 characters)
> Video:           (optional: file name, or "none")
> ```

> [tab] Example
>
> A filled-in template for a three-question interview: question 1 is text only, question 2 has a recorded video prompt, and question 3 shows an image. The files `intro.mp4`, `q2-upset-customer.mp4`, and `q3-churn-chart.png` are attached to the same chat message:
>
> ```text theme={null}
> POSITION
> Name:            Customer Success Manager
> Tags:            US, EMEA
> Intro video:     intro.mp4
>
> QUESTION 1
> Title:           Tell us about yourself
> Description:     Give us a short introduction: your background, what you're working on right now, and why this Customer Success role caught your eye.
> Attachment:      none
>
> QUESTION 2
> Title:           Handling an upset customer
> Description:     Watch the video, where a long-time customer explains why they're frustrated with our product. How would you respond to them, and what would you do in the first 48 hours to win back their trust?
> Attachment:      video: q2-upset-customer.mp4
>
> QUESTION 3
> Title:           What does this chart tell you?
> Description:     This chart shows the monthly churn rate for one of our customer segments over the past year. What stands out to you, and what would you look into first?
> Attachment:      image: q3-churn-chart.png
>
> OUTRO
> Title:           Thanks for completing the interview!
> Text:            We review every interview personally and will get back to you within 5 business days. Questions in the meantime? Reach out to jobs@example.com.
> Video:           none
> ```
>
> Texts can be in any language — your AI tool sends them exactly as you wrote them.

What each part means:

* **Name** — the job title candidates see.
* **Tags** — optional labels to organize your positions in the dashboard.
* **Intro video** — an optional video shown before the first question. There's no intro text: without a video, candidates see the job title, the number of questions, and the **Start Interview** button.
* **Question title and description** — the title is capped at 40 characters, so it's usually a short label (e.g. "Motivation"). Use the **description** (up to 1200 characters) for the actual question, context, or instructions — you'll want one on most questions.
* **Attachment** — optionally, **one** [video prompt](https://docs.hireflix.com/user/building-the-interview/creating-a-video-interview.md) (you asking the question on camera) **or one** [image](https://docs.hireflix.com/user/building-the-interview/image-prompt-questions.md) (a diagram, chart, design, ...) per question. Candidates always answer by recording a video, whatever the attachment.
* **Outro** — the screen candidates see after the last question: a title, a text, and optionally a video.

> [!WARNING]
>
> **Don't tailor question or position settings unless you really need to.** The template has no fields for thinking time, answer time limit, or retakes on purpose: Hireflix's defaults (30 seconds thinking time, 2 minutes to answer, no retakes) work well for most interviews, and your AI tool won't ask about them.
>
> If you do need something different, just say it in plain words:
>
> * **For the whole interview**, e.g. "give candidates 3 minutes per answer" — it sets this once for the entire position, so you can still change it later in one place.
> * **For one specific question**, e.g. "give question 4 five minutes, the rest can keep the defaults" — only do this when a question really needs it. A question with its own settings is marked as having **custom settings** and no longer follows the position-wide values, so later changes to the interview settings won't apply to it.
>
> See [Tailoring the Interview](https://docs.hireflix.com/user/building-the-interview/tailoring-the-interview.md) for what each setting means for candidates.

## Attaching videos and images

Attach files directly to your chat message and reference them by file name in the template (e.g. `video: intro.mp4`, `image: org-chart.png`). Your AI tool uploads them for you. Using the skill in Claude? Make sure you've allowed the websites from the [skill setup](https://docs.hireflix.com/ai/skill.md#setup) first. The MCP doesn't need this.

* **Videos**: max **150 MB** each, any format (Hireflix converts it after upload). A 4:3 aspect ratio fills the screen best; 16:9 works but shows black bars.
* **Images**: max **10 MB** each, any image type (JPEG, PNG, GIF, WebP, SVG). **One image per question** — if you want to show several images, create one image question per image.

> [!NOTE]
>
> Your AI tool never compresses, resizes, or converts your files. If a file is too large or isn't a video/image, it tells you exactly what's wrong (e.g. "that file is 240 MB, 90 MB over the limit") so you can shrink it yourself, for example with [HandBrake](https://handbrake.fr/).

## What gets checked before creating the position

* **Required fields** — a name and at least one question with a title. Fields you leave empty, leave on the placeholder text, or set to "none" are skipped.
* **Character limits** — if a title or text is too long, your AI tool tells you which field and by how much, and asks you for a shorter version. It never shortens or rewrites your text itself.
* **Outro** — if you keep the outro block (even just for the video), it needs both a title and a text.
* **Your content stays yours** — your AI tool doesn't invent questions or texts. Want suggestions? Ask it to draft questions for a role first, then paste the ones you like into the template.

Once the position is created, your AI tool shares its ID. You can review it in the [interview builder](https://docs.hireflix.com/user/building-the-interview/creating-a-video-interview.md), add [branding](https://docs.hireflix.com/user/building-the-interview/branded-interview-page.md), or ask it to make changes, like adding a question or renaming the position.

## What's next?

> [card] [Example prompts](https://docs.hireflix.com/ai/guides/example-prompts.md)
>
> Now that you have a position, invite candidates and ask about their interviews.

> [card] [Tailoring the Interview](https://docs.hireflix.com/user/building-the-interview/tailoring-the-interview.md)
>
> Learn what thinking time, answer limits, and retakes mean for your candidates.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.