Publish-ready content, in your voice.

Content Loop turns briefs, browser captures, and Bananaloop activity into LinkedIn posts, X threads, blog posts, and newsletters — without sounding like a robot wrote them.

Start creating See how it works

No credit card. Bring your own Claude API key.

/dashboard/studio

Brief     → "Our team just shipped runtime policy enforcement for LLM agents."
Voice     → Gorilli
Format    → LinkedIn Post
Tone      → Authoritative

▸ generate

Output ────────────────────────────────────────────
Shipping AI into production without guardrails is
like rolling out a new service with no SLOs.

Today we're shipping runtime policy enforcement
for agent workflows — deterministic checks that
run on every tool call, not just at eval time…

Ships to the channels your audience already reads

LinkedInX / TwitterSubstackGhostMediumBeehiiv

What you get

A studio, not a chatbot

Content Loop is structured around the way real content gets written — capture, context, draft, revise, publish. Every feature is there to shorten the path between an idea and the post.

Voice-faithful generation

Pick a voice (Gorilli, Paolo, or your own), a format, and a tone. Claude writes the draft; you ship it — or tweak a line and ship it.

Browser capture extension

Highlight a paragraph, clip a page, or save a selection from anywhere on the web. Every capture lands in your Ingests queue, ready to become a post.

Multi-format outputs

One brief → LinkedIn post, X thread, long-form blog, or newsletter section. Switch the target format and regenerate in place.

Bananaloop context

Pull your recent activity — tools used, pages read, decisions made — and let the model suggest four content angles grounded in what you actually did this week.

Tone presets

Professional, casual, authoritative, educational, storytelling. Presets are plain-English instructions you can read and override per draft.

Draft library

Every generation is saved automatically. Search, filter by voice or format, publish, or reopen a draft in the studio to continue iterating.

Examples

What a draft looks like on the channel it ships to

Same brief — “We shipped runtime policy enforcement for AI agents” — adapted to four surfaces. Hit a tab to preview.

Gorilli

Runtime governance for AI agents · 12,482 followers

2h ·

Shipping AI into production without guardrails is a bet on luck.

Today we shipped runtime policy enforcement for agent workflows — deterministic checks that run on every tool call, not just at eval time.

What it means in practice:

→ Policies defined once, enforced everywhere
→ Tool calls blocked or flagged in real time
→ Every decision is auditable, with receipts

The goal is boring: make shipping an AI feature feel more like shipping a regular backend service, and less like launching a rocket you hope lands.

#AI #Agents #Governance #LLMOps

👍 ❤️ 🎯 24842 comments · 18 reposts

API first

Drive the studio from anything that speaks HTTP

Every surface — browser extension, dashboard, third-party scripts — hits the same authenticated endpoints.

POST /api/generatecurl

curl https://content-loop.app/api/generate \
  -H "Authorization: Bearer $STUDIO_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "voice": "Gorilli",
    "format": "LinkedIn Post",
    "tone": "authoritative",
    "brief": "We shipped runtime policy enforcement for AI agents."
  }'

# ↳ {
#     "content": "Shipping AI into production without guardrails…",
#     "metadata": {
#       "characterCount": 812,
#       "model": "claude-sonnet-4-6",
#       "usage": { "input_tokens": 412, "output_tokens": 198 },
#       "draftId": "65f0c3d1a9b4e2c0e8f12345"
#     }
#   }

Workflow

From blank page to published in four moves

Capture raw material

Clip pages with the browser extension, drop notes into the Ingests queue, or let Bananaloop pipe your recent activity in automatically.

Shape the brief

Pick a voice, a format, and a tone. Attach an ingest as context or write a one-line brief — whichever is faster. The studio does not care.

Generate and review

Claude writes the first full draft in seconds. Regenerate with a different tone, edit inline, or compare multiple drafts side by side.

Publish or archive

Mark drafts as published when you ship them. Everything stays searchable by voice, format, and source for the next time you need a reference.

Architecture

One pipeline, every surface

Capture sources, a single generation API, and a dashboard that stays out of your way.

Inputs

Browser extension captures
Manual briefs
Bananaloop activity
Pasted URLs & selections

Studio

Voice + format + tone
Angle suggestions
Claude generation
Draft library (MongoDB)

Outputs

LinkedIn posts
X / Twitter threads
Blog posts
Newsletters

Stop staring at a blank editor.

Spin up an account, drop in your first brief, and watch the first draft land in under ten seconds.

Create your account Review the features