Prompt library · card catalog · every card is a feature you are currently paying for
The prompts are the product. Here they all are.
Every 0saas tool ships two to four structured prompts, each mapped one-to-one onto a named SaaS feature it replaces. Each one requests strict JSON against a stated schema, names the free-tier model it was written for, and works with no API key at all if you would rather paste it into a chat model yourself. Copy them, fork them, keep them in a text file — that is the point.
How to run any of these anywhere
- Open the card and press Copy system prompt. In a chat product, paste it as the first message and ask the model to acknowledge it. With an API key, it goes in the
system/systemInstructionfield. - Fill the user template — the
{{double-brace}}slots — and send it. Ask for JSON only; every prompt here already ends with that instruction. - Take the JSON back to the tool page and paste it into the panel's relay box. The tool renders it: a poster on the press bed, slips in the galley margin. No key, no account, no network call from this site.
Free-tier ceilings that apply: Groq's published base is 30 requests/min and 1,000 requests/day on several chat models; OpenRouter's :free variants are 20 requests/min and 50/day under 10 credits purchased; Gemini's free tier publishes per-project limits inside AI Studio rather than as a static table. Sources and read dates: data/allies.json.
Index
| Tool | Prompt | Replaces | Best free ally | v |
|---|---|---|---|---|
| Pressproof Canva alternative |
Layout director | Canva's template picker + Magic Design | Gemini free tier | 1.0.0 |
| Headline press run | Magic Write for social copy (inside the ~20-use allowance) | Groq free tier | 1.0.0 | |
| Reflow to another size | Magic Resize, which is Pro-only at US$144/year | Gemini free tier | 1.0.0 | |
| Galley Grammarly alternative |
Blue-pencil markup | Grammarly's suggestion layer, aimed at what Free cannot do | Gemini free tier | 1.0.0 |
| Tone shift | "Adjust your writing tone" — Unavailable on Free | Groq free tier | 1.0.0 | |
| Reply drafter | The 100-AI-prompts-per-month generation allowance | Gemini free tier | 1.0.0 |
The cards
Loading full prompt text from tools.json…
About these prompts
Why does every prompt demand strict JSON?
Because the tool computes on the answer. A layout spec that is 95% valid JSON draws nothing; a markup whose quotes are paraphrased highlights the wrong words. Prose answers would make every tool a copy-paste chore, so each prompt states its schema and ends by forbidding commentary.
Why do the markup prompts insist on exact quotes instead of offsets?
Character offsets from a language model are unreliable — one silently dropped space shifts every later offset. An exact substring can be verified with a literal search, and when it is not found the tool says so instead of guessing a position. That single rule is the difference between an editor you can trust and one that highlights the wrong word confidently.
Can I use these prompts commercially?
Yes. They are published here to be copied. The repo carries an MIT license. What you may not do is claim that 0saas, Canva, Grammarly or any inference provider endorses you — the same rule this site holds itself to.
Why name a "best" model at all, when models change monthly?
Because "use any model" is not advice, it is a shrug. Each card names the ally and the model id we last verified, the date we verified it, and the reason — usually strict-JSON reliability, context length or request-per-minute headroom. The tool itself lists the models live from your own key, so a stale id in the docs cannot break your run.