Guides

Turn winners into reusable templates

Save your best-performing post structures as templates with placeholders. AI uses them as few-shot examples for higher-quality drafts; you write a 5-word title, AI fills in the rest.

5 min read Beginner

Why content templates compound

Top performers don't write each post from scratch. They have 5-10 proven structures (hook + value + CTA, comparison thread, listicle, case study) and slot fresh details into them.

Mirgent's template system encodes those structures as reusable assets. You save a winner once; AI uses it as a pattern when generating future drafts. Quality compounds; writing time drops.

Scheduling

Save winners as templates with one click

Lead Discovery

Placeholders for variable content

AI Agents

AI uses templates as few-shot examples

Analytics

Track which templates perform best

Build your first template

Three minute setup. Use a recent high-performing post as the seed.

1

Find a winner

Open Analytics → Top Posts (last 30 days). Pick the best-performing one. The structure is what you want to template, not the specific topic.

2

Save as template

Open the post → click 'Save as Template'. Mirgent extracts the structure (hook, body, CTA), prompts you to add placeholders for variable parts.

3

Name + tag

Name it something findable ('B2B Comparison Thread', 'Solo Founder Hook'). Tag with platform + use case so the right template surfaces when you need it.

Tip

Templates with 2-4 placeholders work best. Too few = inflexible; too many = same as writing from scratch. Start with hook + value + CTA placeholders.

Placeholder syntax

Mirgent's templates use {{ variable }} syntax. Variables can be required (must be filled), optional (defaults if empty), or AI-generated (Mirgent fills automatically).

1

{{ topic }}, required

User must supply when applying the template. Surfaces as a form field in the composer.

2

{{ company:Mirgent }}, optional with default

Pre-fills with 'Mirgent' if not overridden. Useful for branded content where the company name is usually constant.

3

{{ ai:hook }}, AI-generated

AI generates the value at draft time using the surrounding context. Most powerful for hooks, CTAs, and platform-specific tweaks.

AI few-shot training

Templates double as few-shot examples for AI agents. When Content Planner drafts a new post, it references your top 3 templates as 'this is what good looks like for this user'.

1

Tag templates with metadata

Mark templates as 'high-performer' (above your engagement median). AI weights these heavier in the few-shot context.

2

Per-platform templates

Twitter templates differ from LinkedIn structurally. Save platform-tagged templates so AI uses Twitter winners for Twitter drafts and LinkedIn winners for LinkedIn drafts.

3

Template-pack rotation

Group templates into packs ('Launch Week', 'Retention Funnel', 'Hiring'). Activate the right pack when you're in that mode; AI focuses few-shot context on relevant winners only.

Template library examples

Five templates we see most-used across the Mirgent community.

Scheduling Guide

Calendar, kanban, and queue modes

AI Agents Guide

Configure autonomous AI workflows

Integrations

Connect RSS feeds and platforms

Lead Discovery Guide

Find leads on Twitter and Reddit

Pitfalls with templates

Template misuse that hurts performance.

  • Using templates verbatim too often. Same hook + same CTA across 20 posts looks templated to the algorithm. Rotate 5+ templates per platform; use each at most weekly.
  • Templates that are too rigid. {{ topic }} alone produces robotic content. Add {{ ai:hook }} and {{ ai:close }} so AI tunes the surrounding language.
  • Saving templates from low-performers. Only template the winners. Below-median posts encode the wrong patterns; AI will reproduce mediocrity.
  • Too many active templates. Above ~15 templates per platform, AI's few-shot context becomes diluted. Curate to top 5-10 per platform; archive the rest.

Template recipes

Three structures that work consistently across Mirgent users.

  1. Hook-Value-CTA (Twitter, 3 placeholders)

    {{ ai:hook }} → 1 specific insight about {{ topic }} → {{ ai:cta }}. Works for daily Twitter cadence; AI handles the rhetorical bits while you supply the topic substance.

  2. Comparison thread (LinkedIn, 5 placeholders)

    Heading: '{{ topic_a }} vs {{ topic_b }}, honest take'. Body: pros of A, pros of B, when each wins, your recommendation. Closes with {{ ai:cta }}. Best for B2B thought leadership.

  3. Listicle with case study (LinkedIn long-form)

    {{ ai:hook }} → '5 things I learned from {{ specific_event }}' → 5 numbered insights, each with concrete number → close with {{ ai:cta }}. Format that drives 3-5x engagement above baseline.

  4. {guide_tmpl_walk4_title}

    {guide_tmpl_walk4_desc}

  5. {guide_tmpl_walk5_title}

    {guide_tmpl_walk5_desc}

Templates FAQ

Can I share templates with my team?

Yes, workspace-scoped. Anyone with workspace access sees and uses the same templates. Cross-workspace sharing via export-import as JSON.

Do templates work with non-English content?

Yes. Save templates in your target language; AI generates content in the same language. Language consistency is at template-level, not workspace-level.

How does Mirgent know which templates to surface?

By platform tag + recency + performance. The composer surfaces top 5 platform-matching templates by your engagement-rate score from analytics.

Can I import templates from other tools?

Manual import only, paste content, mark placeholders, save. We don't have an automated import from Buffer / Hootsuite templates because format conventions differ enough that auto-translation produces broken templates more often than working ones.

What happens to my analytics on templated posts?

Each post tracks both its content metrics AND its template. You can see 'Template X performs 40% above template Y' over time. Use this to retire underperforming templates and double down on winners.

{guide_tmpl_faq6_q}

{guide_tmpl_faq6_a}