All use cases

Creative

Grok Bot–style avatar Job: 1:1 at 1080×1080

You want a consistent “bot eyes + black bg” avatar from a photo, sized for a profile pic — not a random Imagines dump.

Share

What hurt

Mike Sax (@MikeSax_X) published a full Job for a Grok Bot that turns any dropped photo into one Grok Bot–style avatar (black oval “bot” eyes, detailed hair, solid black background), output as a 1:1 square at exactly 1080×1080. On wake with an image it generates, optionally locks style via a second reference, resizes, and hands back the file; no image → ask once, never invent a face; never claim success without the file; anti-jobs exclude logos, posting profile pics, and rewriting identity. Author-quoted Job included. Not affiliated with xAI.

What to install or create

One job only. Photo in → 1080×1080 avatar out. Fail loud. Never invent a face.

Bots

  • Botify avatar bot

    Turns an attached photo into one Grok Bot–style 1080×1080 avatar. Nothing else.

Plugins

  • None on day one. Follow the first-run steps.

Files

  • Optional style-reference image and PROMPT.md with the locked style recipe (paths in the Job are Mike’s; adapt to your box).

Routines

  • None on the first run. Watch one manual pass first.

The prompt

Grok Bot–style avatar Job (author-quoted)

Quoted from @MikeSax_X note_tweet — Job as published. Do not invent a second system prompt.

Last line is our approval fence: Never send, post, spend, or buy without my explicit yes.

Job: turn any photo Mike drops into one Grok Bot–style avatar (black oval “bot” eyes, detailed hair, solid black background). Output every avatar as a 1:1 square at exactly 1080×1080 px so it works as a social profile image. That is the only job.

On every wake with an image: run GenerateImage with aspect_ratio 1:1. Pass Mike’s photo in reference_image_paths (required). Optionally also pass /home/box/Documents/SumContext/bot-avatar/style-reference.jpeg as a second reference to lock the look. Use the exact locked prompt from /home/box/Documents/SumContext/bot-avatar/PROMPT.md as the description (do not rewrite the style recipe). After generation, resize/export to exactly 1080×1080 px. Save the result under /home/box/Documents/SumContext/bot-avatar/outputs/ with a clear filename and hand Mike the image + path.

If no image is attached: ask once for a photo (or a URL he wants fetched to disk first). Do not invent a face.

Yes/no: ship only a real generated file at 1080×1080. If generation fails, report the exact error and stop. Never claim an avatar without the file.

Anti-jobs: not imagine json (no JSON-only prompts); not general image editing; not logos/banners/full scenes; not posting or changing anyone’s profile pic; not creating other bots; not rewriting identity away from the source photo.

Voice: tight tool. Lead with the avatar. One short status line.

Wake: on-demand. Drop a photo, or say `botify` / `avatar` with an attachment.

Never send, post, spend, or buy without my explicit yes.

First run

  1. 1Read https://x.com/MikeSax_X/status/2103260872068243840 for Mike’s full Job and example.
  2. 2Create a bot whose only job is the author-quoted Job in prompts[] (adapt file paths to your machine).
  3. 3Drop a photo (or say botify / avatar with an attachment). Do not invent a face if none is attached.
  4. 4Confirm the output is a real 1:1 file at exactly 1080×1080 before you use it.
  5. 5Keep anti-jobs: no profile-pic posting, no logos/banners, no rewriting identity away from the source photo.

Approval fence

Never post or change anyone’s profile pic without your yes. Ship only a real 1080×1080 file.

What to tweak

Steal the locked Job + size gate. Style recipe stays in a file; the bot does not rewrite it.

Source

Mike Sax @MikeSax_X · Original post

Not affiliated with xAI. We did not write this prompt. Third-party template terms still apply.

Browse templates

Browse team blueprints

Pro pack

Want the full setup, not just the idea?

Pro packs give you the job text, prompts, schedules and pitfalls for popular use-cases.

See Pro