Skip to content



Behind the Scenes of the Cognyn Planner Beta

How Cognyn turns a plain-English goal into a multi-AI plan — step by step.



🧭 The Planner’s Job

The Cognyn Planner Beta is a reasoning layer. You describe what you want — for example,
“Create a campaign for my charity” — and the planner responds with an ordered workflow that coordinates
multiple AIs (text, image, video, email, CMS). In this beta, the focus is on planning;
full execution comes next.

Under the hood, the planner does four things: parse the goal, map tools,
chain steps, and estimate cost/effort. Let’s walk through each part.



1️⃣ Goal Parsing

The planner first extracts the intent (what outcome is desired),
the domains involved (copywriting, design, video, email, web),
and any constraints (tone, audience, deadlines, assets).
It uses lightweight rules and semantic patterns to identify these from natural language.

Examples it looks for include:

  • “launch,” “campaign,” “poster,” “email,” “landing page,” “donation,” “event,” “deadline.”
  • “assets provided” vs. “assets needed.”



2️⃣ Tool Mapping

Next, the planner chooses the best capability providers for each part of the job.
For example: OpenAI (text), Midjourney (images),
Runway (video), Mailchimp (email),
and WordPress (CMS).

It builds a candidate set based on the domains detected in parsing. In the beta, we present these as
“suggested tools,” but the architecture will allow users to substitute or lock in preferred vendors
later (e.g., “use Leonardo instead of Midjourney”).



3️⃣ Workflow Chaining

With tools selected, Cognyn sequences them into a logical chain of steps.
Output from one becomes input to the next — for example, copy → image prompts → video script → email body → landing page.
This is where the “project manager for AIs” metaphor comes alive.

  • Text → brand story, tagline, body copy
  • Images → posters or social graphics driven by text
  • Video → short promo based on script and key visuals
  • Email → subject line, body, and call-to-action
  • CMS → publish landing page with all assets assembled



4️⃣ Quality Checks & Iteration

Even in planning mode, Cognyn anticipates feedback loops: “If the poster style is off-brand,
regenerate with a calmer palette,” or “If subject lines score low, try three alternatives.”

Execution mode will formalize these loops using evaluation prompts and heuristics
(readability, tone, brand consistency). These micro-iterations add up to big gains —
fewer dead ends, higher quality results, and faster publishing cycles.



💠 Credit & Cost Estimation

The beta introduces an Estimated Credits field — a friendly shorthand for
time, complexity, and third-party usage. Different tools have different weights:
a video render “costs” more than an email draft.

In future versions, credits will correspond directly to real-world provider usage —
tokens, renders, API calls — creating transparent cost-awareness at every step.



🔒 Privacy by Design

Cognyn is orchestration-first and user-owned.
Your inputs, outputs, and API keys remain yours.
We never train proprietary models on your data without consent,
and only the minimum context required is sent to external tools.

This architecture is privacy-forward by design. For details,
see the Privacy Policy.



🚀 What’s Coming Next

  • Live Orchestration — execute plans through real connectors (OpenAI, Midjourney, Runway, Mailchimp, WordPress).
  • Editable Pipelines — drag to reorder steps, swap tools, and add review gates.
  • Templates — one-click flows like “Product Launch,” “Event Drive,” “Fundraiser,” and “Course Promo.”
  • Team Mode — shared workflows, roles, and audit trails.
  • Developer API — define custom tools and evaluators.

Try it yourself — visit the Cognyn Planner (Beta),
describe a goal, and watch the workflow appear.
Want early access to live orchestration? Join the
Early Access List.



Cognyn — The Mind That Connects Minds.