Pip Has One Job: three stickman episodes for 40 credits each
A three-episode AI stickman comedy made from still images: 40 credits per video, one failed try, and early view counts with read ages. Run 2026-10-27; numbers read up to 2026-10-28.
Format: Stickman comedy from still images with narration, three episodes of about half a minute each, vertical 9:16. Everything shown in the videos is AI-generated. Numbers are from our own accounts and are small; each carries the date it was read, because platform numbers change as a video ages. This page makes no promise of views.
What we made
Pip Has One Job is a three-episode stickman comedy about procrastination: Pip has one job, washing one cup, and it takes nine days. It is built from six generated still images per episode with a narration voice and word-by-word captions, not from animated clips. Everything is AI-generated and labelled.
We used the STICKMAN art style and one series, so the same stick figure appears in every episode. Each episode was one create call with a script of six short scenes, each with its own picture prompt. The first try of episode 3 showed Pip drinking from the mug where the story needed her to hold it out, so it failed our check and was not posted; the retry rewrote the picture prompts ("holding a mug out in front of her at arm's length") and passed.
One thing MCP users should know from this run: the series id is what pins the stick figure across episodes. The chat video and draft tools do not take a series id, so episodes made there get a per-video look. We read that from the code and did not A/B it.
Credits and cost
| What | Number | As of | Source file : line |
|---|---|---|---|
| Each video: six still images plus voiceover, no clips | 40 credits per video | run 2026-10-27 | qa.md:9 |
| Four videos made (three episodes plus one failed first try of episode 3) | 4 videos, 160 credits | run 2026-10-27 | qa.md:14 |
| Estimated provider cost for all four (list prices, not an invoice) | about 0.08 USD | run 2026-10-27 | qa.md:14 |
| Length of the three posted episodes | 25.77 s, 30.37 s and 26.63 s | run 2026-10-27 | qa.md:20-23 |
Total credits (four videos at 40 credits each): 160 credits, which is $3.20 at the $0.02-per-credit list rate. Plan prices differ; see what things cost.
Credits are what the product charged for the four videos. The provider cost is our estimate from list prices; the product itself records zero for the image provider, so it is not an invoice.
What the platforms showed so far
| What | Number | As of | Source file : line |
|---|---|---|---|
| Our own stop rule, written before posting | stop and change the hook if the median of 3 episodes is under 300 views, or "chosen to view" is under 40% | rule written 2026-10-27 | posts.md:53 |
| YouTube, episode 1: views | 3 views, 2 engaged | 2026-10-28 11:30-11:45 ICT, 5.3 h after posting | baseline-youtube.md:56 |
| YouTube, episode 2: views | 3 views, 2 engaged | 2026-10-28 11:30-11:45 ICT, 2.2 h after posting | baseline-youtube.md:57 |
| Facebook Page, episode 1: reach and views | reach 298, 346 views | 2026-10-28, read 06:40-09:40 ICT, 2.6 h after posting | measurements-2026-10-28-autoshort.md:26 |
| Facebook Page, episode 1: later read | 416 views, 379 viewers, 3-second views 170 (40.9%), average watch 5 s | 2026-10-28 about 11:20 ICT, 5.5 h after posting | baseline-facebook.md:41 |
| Instagram: episode 1 and episode 2 views at read time | 33 views and 2 views | 2026-10-28 11:27 ICT, about 5 h and 2 h after posting | baseline-instagram-threads.md:17 |
| X, episode 1: views | 5 views | 2026-10-28 about 11:35 ICT, 4 h after posting | baseline-x.md:22 |
| TikTok, episode 1 | 0 views | 2026-10-28 11:28 ICT | baseline-tiktok.md:37 |
| LinkedIn (profile), episode 1: impressions | 28 impressions | 2026-10-28, read 06:40-09:40 ICT, 2.5 h after posting | measurements-2026-10-28-autoshort.md:32 |
These reads are hours old, not the 24-hour checkpoint our stop rule asks for, so we draw no verdict. Facebook is the only surface with a sizeable number. Own views are not excluded. Numbers on every platform keep moving as the reels age.
What we learned
- Still-image episodes are the cheapest format we ran: a fixed 40 credits a video, with no clip-by-clip cost.
- Word-based scene timing runs ahead of the voice, because the scene lengths are computed from word counts and the voice is slightly faster. It stayed inside our tolerance, but you can see the picture change early.
- Prompt wording matters more than anything: "holding a mug" was drawn as drinking. "At arm's length, far from her face" fixed it.
- The stick figure drifts a little even with a locked description (blush marks appeared on some frames).
What we would do differently
- Wait for the 24-hour read before judging, as our own rule says.
- Post episodes on one channel in order, on the same day, so a viewer can find the next one.
- Use the series id from the first episode, so the look stays pinned.
The public videos
Links as recorded in our publishing ledger. Platforms can remove or restrict a video later.
Try the same kind of prompt
Start from a script you can read and edit first. A drafted script is stored word for word and costs nothing; the render is where credits are charged.
Open the prompt libraryConnect AutoShort to Claude, ChatGPT, Cursor or Gemini CLIFree script preview
A prompt does not make a video perform. It makes the script and the cost visible before you spend credits. Results of ours are above.
Data sources
Files in our internal research folders that every number on this page comes from:
- growth-ops/research/stickman-series-2026-10-27/qa.md, posts.md, howto.md
- growth-ops/research/formula-batch-1-2026-10-28/baseline-youtube.md, baseline-facebook.md, baseline-instagram-threads.md, baseline-x.md, baseline-tiktok.md
- growth-ops/research/measurements-2026-10-28-autoshort.md
- growth-ops/ledger/published.md (public links)