Rook Intelligence

The quiet engine
behind every word.

Trained for short form reply. Tuned to sound like you. Ready to run on the model you trust.

Rook IntelligenceDrafts
Today's briefing

Seven new posts from your fifteen accounts. Five drafts ready. Naval and April are the highest signal posts this morning. Shaan posted twice in eighteen hours.

Naval RavikantPre generated
The real 10x engineer is the one who knows which problems to ignore.
This is the part most teams miss. Priority is not about working faster. It is about refusing the wrong work.
April DunfordPre generated
Positioning is not a tagline. It is the reason your buyer says yes.
Agreed. The strongest positioning reads like a calm sentence. No jargon. Just a buyer reason.
A different kind of layer

Not a chatbot. A writing layer.

Rook Intelligence is the prompt, rule, and persona layer that sits above your model of choice.

Short form

Tuned for replies, not essays.

Most models want to impress you with length. Rook Intelligence is trained to stop early. Two sentences, three sentences, maybe four. The cadence of a good LinkedIn comment, not a blog post.

  • Length aware
    Drafts land between fifteen and sixty words on most posts.
  • Post aware
    It reads the original post first. Every draft answers that post, not a generic template.
  • Tone aware
    It reads your tone before it writes. No mode switching.
Rook IntelligenceDrafts
Today's briefing

Seven new posts from your fifteen accounts. Five drafts ready. Naval and April are the highest signal posts this morning. Shaan posted twice in eighteen hours.

Naval RavikantPre generated
The real 10x engineer is the one who knows which problems to ignore.
This is the part most teams miss. Priority is not about working faster. It is about refusing the wrong work.
April DunfordPre generated
Positioning is not a tagline. It is the reason your buyer says yes.
Agreed. The strongest positioning reads like a calm sentence. No jargon. Just a buyer reason.
Model choice

Pick the brain behind the voice.

Anthropic. OpenAI. Google. Groq. xAI. Plug the model you trust into the workspace and Rook Intelligence adapts its prompt shape to that model.

  • Five providers
    All major labs, all on one settings page.
  • Bring your own key
    Encrypted at rest. Never returned in full to the browser.
  • Per workspace
    Different personas, different models. Your call.
AuthenticationAES 256 GCM
Bring your own key
Keys are encrypted at rest and never returned in full to the browser.
ProviderModelKeyStatus
AnthropicClaude Opus 4.7sk-ant-api03 ████████ JX9k Active
OpenAIGPT 4.1sk proj ████████ 2F7t Ready
GoogleGemini 2.5 ProAIza ████████ Qe4s Ready
xAIGrok 4xai ████████ 81pD Ready
GroqLlama 3.3 70Bgsk ████████ nT2v Ready
What it does

Writing, not generation.

The difference is care. Rook Intelligence is asked to think, not to fill the box.

Draft, not paste

Every draft is a first pass, shaped by your persona and the post. Editable, never locked.

Persona aware

Samples and rules are injected into every request. No generic baseline slips through.

Model agnostic

Swap providers without rewriting a single sample. The persona is yours, not theirs.

Fast

Drafts come back in seconds during tracking, and in under three when you hit regenerate.

Private prompts

Your samples and rules stay in your workspace. Never pooled, never trained on.

Per workspace

Run Anthropic for yourself and Groq for your brand. Both are one click apart.

5
Providers supported
2s
Average regen time
0
Model lock in
Switch any time.
100%
Of prompts private
Never pooled, never trained on.
Rook Intelligence

The brain you choose. The voice you keep.

Open a workspace and plug in your key. Rook Intelligence takes it from there.