/PromptTask
Turn a rough idea into a precise prompt for a task — e.g. extracting invoice totals from scanned PDFs.
Your AI Command Vault
Summoning the AI overlords…
Search
220 commands for “prompt”
Naming the exact shape of the answer removes most rewriting.
Asking for clarifying questions before an answer removes guesswork.
Try two anonymous models on the same prompt and vote for the better answer.
Browser workspace for prompting Gemini models and testing an API key.
Prompt techniques and workarounds shared by heavy assistant users.
Craft discussion plus a wiki of free tools, prompts and publishing advice.
Turn a rough idea into a precise prompt for a task — e.g. extracting invoice totals from scanned PDFs.
Turn a rough idea into a precise prompt for an agent — e.g. a customer-support agent that looks up order status via API.
Rewrite a rough prompt into a precise, well-constrained instruction.
Turn a rough idea into a precise prompt for a chatbot — e.g. a website FAQ chatbot for a SaaS billing product.
Turn a rough idea into a precise prompt for a summarization job — e.g. daily summarization of a 200-message Slack channel.
Turn a rough idea into a precise prompt for a creative brief — e.g. a brief for a whimsical mascot illustration for a coffee.
Turn a rough idea into a precise prompt for an extraction job — e.g. pulling names and dates from handwritten meeting notes.
Find cost waste and savings in a prompt — e.g. a 200-word support-ticket triage prompt used in production.
Run a structured quality audit of a prompt — e.g. a 200-word support-ticket triage prompt used in production.
Stress-test the weakest points of a prompt — e.g. a 200-word support-ticket triage prompt used in production.
Build a scoring rubric for a prompt — e.g. a 200-word support-ticket triage prompt used in production.
Trace and reduce latency in a prompt — e.g. a 200-word support-ticket triage prompt used in production.
Propose the three highest-leverage improvements to a prompt — e.g. a 200-word support-ticket triage prompt used in production.
Diagnose why the model is misbehaving on a retrieval-augmented generation prompt — e.g. a support-doc QA prompt that must cite the source paragraph.
Write a high-quality prompt for a retrieval-augmented generation prompt — e.g. a support-doc QA prompt that must cite the source paragraph.
Define fair benchmarks and baselines for a prompt — e.g. a 200-word support-ticket triage prompt used in production.
Tighten up and remove ambiguity from a retrieval-augmented generation prompt — e.g. a support-doc QA prompt that must cite the source paragraph.
Hunt for regressions introduced in a prompt — e.g. a 200-word support-ticket triage prompt used in production.
Flag inconsistencies across a prompt — e.g. a 200-word support-ticket triage prompt used in production.
Write a high-quality prompt for an assistant's system prompt — e.g. a customer-support bot that must stay on-topic and never quote.
Shrink the token footprint of a retrieval-augmented generation prompt — e.g. a support-doc QA prompt that must cite the source paragraph.
Turn into a reusable template with variables a retrieval-augmented generation prompt — e.g. a support-doc QA prompt that must cite the source paragraph.
Diagnose why the model is misbehaving on an assistant's system prompt — e.g. a customer-support bot that must stay on-topic and never quote.
Set up prompt/model versioning for a prompt/response caching layer — e.g. caching identical FAQ answers for one hour.