Getting started

Take your startup from context to concrete work.

Lipsius is an AI workspace for founders. Give it useful context, review the first read, then steer the Agent toward a specific result you can inspect and improve.

Already have a workspace? Log in to Lipsius.

Context First files Startup read Directed work Business check

Keep your place

Five-step checklist

0 of 5 complete
Getting-started tasks

Start

Create your Lipsius workspace

Go to app.lipsius.io/signup and create an account. If you already have one, use the login page. Your workspace is where startup context, generated files and ongoing work stay together.

Good practice

Describe the business clearly, but do not paste passwords, API keys, banking details or other credentials into startup context or Chat.

Ground the work

Add useful context, then create the first files

Specific context makes the first output more useful. Include the product, stage, customer, problem, current evidence, near-term outcome and real constraints. After onboarding, choose Create first files so Lipsius can turn that context into a starting set of workspace documents.

Detailed example context

A founder testing demand before building

Product: LedgerLoop, a weekly cash-flow briefing for small creative agencies.
Stage: Idea. No product or paying customers yet.
Customer: Agency owners with 5–20 employees who manage cash flow in spreadsheets.
Problem: They can see account balances, but struggle to spot late invoices and upcoming shortfalls early enough to act.
Current evidence: Three informal conversations; two owners said Friday cash reviews take more than an hour. This is directional, not validated demand.
30-day outcome: Interview 10 qualified agency owners and get 3 to try a manual weekly briefing.
Constraints: One founder, 10 hours per week, €300 research budget. Do not build software until the manual workflow is tested.
Open question: Is a weekly decision-ready briefing valuable enough to replace the owner's current spreadsheet routine?

Review

Open the startup read in Agent

In Agent, open the startup read and inspect the first files. Treat them as a structured starting point—not a verdict. Check whether the customer, problem, evidence and riskiest assumptions match what you actually know.

  • Keep facts and assumptions separate. A plausible statement is not customer evidence.
  • Correct the workspace early. Tell Chat what is wrong, missing or outdated.
  • Choose the next decision. Prefer work that reduces uncertainty over more generic documents.

Direct the Agent

Ask Chat for a concrete desired outcome

Strong requests name the context to use, the deliverable, constraints, where to save it and how to verify it. Start with one bounded outcome; review what changed before expanding the scope.

Validation plan

Using my startup context and existing workspace files, create a seven-day demand-validation plan for the customer segment already defined. Include five interview questions that avoid leading the participant, a recruiting message, daily tasks capped at 45 minutes, and a decision rule for continue, change or stop. Save it to plans/seven-day-validation.md. Label assumptions clearly and do not invent customer evidence.

Risk review

Review the startup read and assumption log. Identify the single assumption that could invalidate the idea fastest, explain why using only evidence already in the workspace, and design the smallest test I can run this week for under €100. Save a one-page brief to experiments/riskiest-assumption-test.md with a pass threshold, fail threshold and what I should do for each result.

Build, but do not ship

Build a mobile-friendly static interview landing page from the approved value proposition in my workspace. Use plain HTML, CSS and JavaScript with no external dependencies or data-collection form. Save it under builds/interview-page/, run syntax and content checks, and report the exact results. Do not deploy it; prepare a release checklist for my review.

Before anything leaves the workspace

Check Business before deployment, payments or promotion

Research, drafting, standard workspace files and code execution can happen inside the workspace. External actions depend on your connected services and the permissions you have explicitly enabled. Open Business to review readiness before asking the Agent to act.

ActionWhat must be readyHonest expectation
DeployAn eligible managed-static pilot or a connected deployment providerManaged static hosting is currently an owner pilot. Other deployments require your connected provider.
Collect paymentsYour merchant Stripe setup completedA payment offer or link is not proof that a customer paid.
Publish sociallyYour social account connected and posting permission enabledYou approve the access and cadence; outcomes such as followers or sales are never guaranteed.
Run adsYour ad account, locations, creative, duration, enabled permission and explicit budgetPaid ads spend from your platform account. They are not no-cost, and performance is not guaranteed.
You stay in control

Lipsius works within saved account connections, permissions and budgets. If a required service is not connected or an action is not enabled, that external step cannot be completed.

One useful outcome at a time

Start with context. Finish with something you can inspect.

Create a workspace, review the first read and give the Agent one clear job.

Get started with Lipsius