Postdom

Agents could make the plan. They still could not safely run it.

AI agents can write captions and organise a content calendar, but publishing is where the workflow breaks. Each platform has its own authorisation, timing rules, failure states and measurement gaps. A chat response is not enough when work has to run hours later and every destination needs a trustworthy outcome.

Postdom was a client engagement spanning the proposition, product strategy, UX, technical delivery, pricing and launch. We used AI throughout the work while keeping the product grounded in the operational details that make automation dependable.

Design the service for the agent, and the controls for the person.

Postdom gives AI agents one API and a set of MCP tools to plan, schedule, publish and measure short-form video. The agent works through a machine-readable contract; the person sets the boundaries, approves the plan and completes social authorisation on the platform's own OAuth screen.

The same product supports direct agent use, embedded publishing inside another SaaS product and separate client workspaces for agencies.

• Agent-readable API and MCP tools
• Plan-level human approval
• Durable scheduling
• Per-account autonomy controls
• Explicit publish outcomes
• Measurement availability states

Make autonomy useful without making it opaque.

Take the product all the way to something a client could sell and support.

The work moved from the core proposition into the user journeys, agent contract, account model, design system, publishing workflow, measurement model, pricing and public launch. Product and engineering decisions were made together because the customer promise depends on the underlying execution being durable.

The result is a self-serve product with free and paid plans, agent keys, connected social accounts and publishing support across TikTok, Instagram Reels and YouTube Shorts.