Prompt template library for content ops (free, copy-ready)

This page is a prompt template library for content teams: reusable, copy-ready prompts (not a Notion dump of random ChatGPT one-liners). Use it when search intent is prompt library free, AI prompt template library, or ChatGPT prompt library for shipping briefs → drafts → SEO — not image-gen packs.

Free, no signup: copy the blocks below into any chat or API. Pair with the AI content ops workflow, prompt formatter & cleaner, and production system prompts (system vs user).

Content ops fails when every writer invents a new chat ritual. A shared library fixes that: same inputs, same gates, same quality bar.

How to use this library

1. Keep a short style sheet (tone, banned phrases, product names) in a shared doc.
2. Fill every {{variable}} before sending. Incomplete variables produce vague drafts.
3. Run one stage at a time. Do not ask for outline + full draft + SEO meta in one shot.
4. Store the filled prompt next to the brief in git or Notion — chat history is not the archive.

Estimate token spend with the token estimator before you batch a week of drafts.

Template A — Intake brief expander

Use when a stakeholder dumps a half-sentence idea and you need a real brief.

You are a content ops editor. Expand the rough idea into a structured brief.
Do not write the article. Output Markdown with these exact headings:

## Audience
## Job to be done
## Primary search intent
## Must-include facts (with source placeholders)
## Must-avoid claims
## Format and length budget
## CTA
## Open questions for the requester

Rough idea:
{{idea}}

Product / brand context:
{{context}}

Template B — Outline from frozen brief

You are a senior SEO editor. Produce an H2/H3 outline only — no prose paragraphs.
Align structure to search intent and the brief. Flag any missing facts as [NEED FACT].

Brief:
{{brief}}

Primary keyword: {{keyword}}
Secondary topics: {{topics}}

Human gate: edit structure, freeze outline, then move on.

Template C — Section draft

Long pieces drift when you generate everything at once. Draft one section at a time.

Write the section titled "{{h2}}" for a practical guide.
Audience: {{audience}}
Tone: {{tone}}
Constraints:
- No fake personal testing claims ("I tested", "in our lab")
- Cite only facts provided below; otherwise write [NEED SOURCE]
- Target {{word_count}} words for this section
- End with a short checklist or actionable next step if natural

Frozen outline (for context):
{{outline}}

Facts for this section:
{{facts}}

Style rules:
{{style_sheet}}

Template D — Fact & claim audit

You are a skeptical fact-checker. Review the draft. List every claim that needs a source,
every number that looks estimated, and every absolute ("always", "never", "best").
Do not rewrite the article. Output a table:

| Claim | Risk | Suggested fix |

Draft:
{{draft}}

Template E — SEO pack (after human edit)

Given the final edited article and primary keyword, produce:
1. Title tag options (3), ≤60 characters
2. Meta description (2), ≤155 characters
3. Slug suggestion
4. H1 confirmation (or one alternative)
5. Internal link suggestions using ONLY this allow-list: {{internal_urls}}
6. FAQ of 3 questions answered in 2 sentences each from the article

Do not invent product features. Keyword: {{keyword}}

Article:
{{final_article}}

Template F — Social / newsletter cutdown

Turn the article into:
- 1 LinkedIn post (≤1,200 characters), practical tone, no hype
- 1 X/Twitter thread outline (5 tweets, bullets only)
- 1 email subject + 80-word blurb

Article summary:
{{summary}}
CTA URL: {{url}}

Template G — Style normalization pass

Rewrite for clarity only. Preserve meaning and claims. Apply style sheet.
Remove filler openers ("In today's world…"). Keep technical terms.
Return the full revised draft.

Style sheet:
{{style_sheet}}

Draft:
{{draft}}

FAQ — prompt library vs random prompt dumps

Is this a free prompt library I can use without signup?
Yes. Everything on this page is copy-paste in the browser. No account, no download wall.

How is a prompt template library different from a ChatGPT prompt dump?
A library for content ops is staged (brief → outline → section → audit → SEO) with variables and human gates. Random “100 prompts” lists usually skip ownership, sources, and versioning.

Should these live in system or user messages?
Put durable role/refusals/output contracts in the system channel; put today’s brief and facts in the user message. See system prompt vs user prompt.

Where do Midjourney / image prompt libraries fit?
They don’t — different SERP. This library is for editorial/content-ops text workflows only.

Operating rules for teams

• Version the templates. Put them in a repo (prompts/content-ops/v3.md). Bumping a version is cheaper than arguing in Slack.
• Separate system vs user. If you use an API, put role and hard constraints in the system message; put brief and facts in the user message. See production system prompts.
• Budget tokens. Section drafts + audits use more input tokens than people expect. Track weekly with a shared rate table.
• Never skip the human gate on claims, legal, pricing, or medical/financial advice.

Common failure modes

Failure	Fix
One mega-prompt for everything	Split stages; freeze outline
Style rules buried in chat	Dedicated style sheet doc
Model invents case studies	Ban “I tested” language; require sources
SEO meta written before edit	SEO pack only on final copy
Writers paste competitor dumps	Clean with prompt cleaner first

Minimal weekly cadence

Monday: fill briefs. Tuesday: outlines + freeze. Wednesday–Thursday: section drafts + audits. Friday: SEO pack + publish. Measure cycle time, not “AI vibes.”

If you need a full pipeline diagram, read building an AI content pipeline. For comparing writing models, see ChatGPT vs Claude vs Gemini for writing.