CO-STAR Prompt Framework: Complete Guide with Examples (2026)
Last updated June 28, 2026 · 2 examples · Works with ChatGPT, Claude, Gemini
Quick Answer
CO-STAR is a six-part prompt engineering framework developed by the Singapore Government Technology Agency: Context sets the background, Objective defines the goal, Style specifies writing approach, Tone controls voice, Audience identifies who will read the output, and Response defines the format. It is the best framework when audience fit and communication style are the primary concerns.
What Is the CO-STAR Prompt Framework?
CO-STAR was developed by the Singapore Government Technology Agency (GovTech) and won first place in Singapore's GPT-4 prompt engineering competition in 2023. It is uniquely focused on the communication layer of a prompt — not just what to produce, but for whom, in what voice, and in what format.
What Does CO-STAR Stand For?
Best for content writing where audience and tone are critical.
C
Context
Background information that frames the task. What is the situation, product, or topic the AI needs to understand before writing?
O
Objective
The specific goal of the output. What should the reader think, feel, or do after consuming this content?
S
Style
The writing style — analytical, narrative, persuasive, instructional, conversational. Can reference a specific author or publication.
T
Tone
The emotional register — professional, friendly, urgent, empathetic, authoritative. More granular than style.
A
Audience
Who will read the output. The more specific the better — not 'developers' but 'senior backend engineers who are skeptical of new tools'.
R
Response
The output format — length, structure, headers, bullet points, word count, language.
+R
Rules (optional)
Cross-cutting constraints added by Promptary as an optional seventh field.
When to Use CO-STAR
Use CO-STAR for blog posts, marketing copy, emails, social media content, documentation, and any task where the audience's understanding and the writer's voice matter as much as the content itself. For tasks where process control matters more than style, use RISEN.
CO-STAR Examples
Context: Promptary has just launched its Developer plan which gives developers a REST API endpoint for every saved AI prompt.
Objective: Get the reader to upgrade to the Developer plan within 24 hours by showing them exactly how the API saves time.
Style: Direct and technical — like a developer writing to another developer, not a marketer writing to a lead.
Tone: Excited but not hype-y. Confident. Peer-to-peer.
Audience: Indie developers and small startup teams who already use AI in their products and are familiar with REST APIs.
Response: Email with subject line. Under 200 words. No bullet lists. One CTA link at the end.
Context: I shipped a feature that gives AI prompts stable REST API endpoints so they can be updated without redeploying code.
Objective: Build awareness of the problem and position Promptary as the solution, driving profile visits and app signups.
Style: Storytelling — start with the problem I personally experienced, then the solution.
Tone: Authentic founder voice. Honest about what was hard. Not corporate.
Audience: Developers and technical founders who build AI-powered products.
Response: LinkedIn post. 150–200 words. Hook in the first line. No hashtags. End with a soft CTA.
CO-STAR vs Other Frameworks
CO-STAR vs RACE: RACE is better for execution tasks. CO-STAR is better when how you say something matters as much as what you say. CO-STAR vs CRAFT: CRAFT adds an Action field and omits Audience. Use CO-STAR when targeting a specific audience is the primary concern.
Use this CO-STAR template in Promptary — free
Save your first CO-STAR prompt in the structured editor and get a permanent REST API endpoint. Personal plan is free forever, no credit card required.
Start free
Frequently Asked Questions
What does CO-STAR stand for?
CO-STAR stands for Context, Objective, Style, Tone, Audience, and Response. It was developed by the Singapore Government Technology Agency (GovTech) and won first place in Singapore's national GPT-4 prompt engineering competition.
Is CO-STAR better than RACE?
Neither is universally better. CO-STAR outperforms RACE on content writing tasks where audience and voice matter. RACE outperforms CO-STAR on analytical, technical, and execution tasks.
What is the difference between Style and Tone in CO-STAR?
Style is the structural approach — analytical, narrative, instructional. Tone is the emotional register — professional, warm, urgent. A piece can be analytical in style with a warm tone.
Can CO-STAR be used for image generation?
CO-STAR is designed for language output. For image generation, use Promptary's Image Gen framework.