Opportunities › Multi-agent AI workspace for complex projects › Proposed app
Tracework
A block canvas where AI agents do project work you can inspect, edit and branch
- Score
- 44
- confidence 45
- Evidence
- 6
- items on the page
- Rivals
- 3
- named in sources
- First version
- 45 d
- estimate
- Type
- Desktop app
- PWA: runs in the browser, installs on Mac and Windows
- Market since
- 2022
- Emerging market · AI estimate
ScreensConcept
Desktop app: Founders manage complex multi-block projects with parallel agents, branch strategies and deliverables during focused planning and review work. Sample screens on a computer, in the proposed brand. Their numbers are sample data.
New project
Draft ready to review7 blocks
About this appAI proposal
Tracework turns a complex project into a canvas of blocks: research, analysis, model, deck. Each block holds its own result. Parallel agents fill the blocks, and you can see what feeds what.
Change one block and rerun only it and the blocks downstream. Branch from any block to try a second strategy from the same upstream context. Every conclusion links back to the blocks behind it.
Made for founders and product managers running competitive analysis, financial models, SEO audits or pitch decks who have outgrown a single chat thread. A chat bar can draft and connect blocks for you, so there is no interface to learn first.
- Tagline
- See the work, change one step, keep the rest.
- Voice
- ClearPracticalCalm
- Why this name
- Trace points to the audit trail; work is what the agents do. Two plain words joined, easy to say.
In the first version
Members say which features they would use. It tells PWAS what people want; it changes no score.
Blocks with visible results
Each step is a block with its own input, output and notes. Intermediate work stays on the canvas instead of inside a model's memory.
No member yet
Rerun one block
Edit a block and rerun only it and what depends on it. Everything upstream stays as it was.
No member yet
Branch from any block
Fork a block to test another strategy. Branches share the upstream context and sit side by side.
No member yet
Parallel agents
A goal is split into subtasks that agents work on at the same time, each writing to its own block.
No member yet
Brand identityAI proposal
Check the name against trademark registers before using it.
Palette
#F6F2EAPaper · surface#1F2A2EInk · text#14595CDeep Teal · primary#E0902BAmber · accent#8FB3A3Sage · data
Typography
Aa
Tracework
ABCDEFGHIJKLMNOPQRSTUVWXYZ 0123456789
See the work, change one step, keep the rest.
PricingAI proposal
The material gives no competitor prices and no monetization. These tiers are a starting point to test; real model costs per run must set the limits.
Free
Free
- 1 project
- Limited agent runs
- Audit trail
Pro
$29 a month
- Unlimited projects
- Branching and partial reruns
- Long-running jobs
- Export deliverables
Team
$79 a month
- Up to 5 seats
- Shared canvases
- Higher agent run limits
Would you pay Pro · $29 a month?
Your answer tells PWAS which prices people accept. It changes no score.
No answers yet.
What it could be worthEstimate
What a software business making this much sells for today, from real deals. Pick a monthly revenue.
If it makes
Worth about
$194,000
Sale range $132,000 to $263,000
$120,000 a year in revenue
From 87 software businesses listed and sold on Empire Flippers in the last year, revenue vetted by the broker: they sell for about 19× their monthly revenue, most between 13× and 26×.
Measured on profit, the middle deal is 32× monthly net profit: about 2.7 years of profit.
An estimate, not an offer or financial advice. A real price depends on profit, growth, churn, how much of the work is the owner's, and the buyer.
How it would growAI proposal
How the product would reach new customers through its own use, and how the first ones would be found. A plan to test, not a result.
- Built-in loop
- Shared outputs. Users share read-only canvas links or exported deliverables (decks, PRDs, analyses) with co-founders, investors and teammates, who can open the audit trail behind each conclusion.
- Do they become customers?
- Viewers are often founders or PMs who could run their own projects. Test whether read-only viewers click through to start a project, and whether shared links lead to signups.
- First customers
- Solo founders and early-stage product managers at startups under 20 people who prepare competitive analyses and pitch decks.
- Where to reach them
- Hacker News: Post a Show HN with a sample canvas, since the source discussion drew 109 points and 69 comments.
- Founder communities: Share example canvases in startup forums and newsletters, with offers to build one live.
- Direct outreach: Offer a free first project to 30 founders and PMs, then ask for feedback.
- First milestone
- Ten paying founders or PMs on Pro within 60 days, with at least half running a second project.
Why people already pay for this
- VerifiedHacker News post received 109 points and 69 comments
- Public claimChat is a linear thread, making it impossible to see how AI models connect pieces together, correct one step without rerunning everything, or branch off to explore multiple strategies side by side.
- Public claimUsers of early versions of the tool appreciated being able to branch from a starting point and generate multiple deliverables (prototype, PRD, competitive critique, pitch deck) while sharing upstream context.
- Public claimNew users wanted the tool to replicate manual workflow usage through a chat layer that would automatically generate and connect blocks, rather than learning the interface.
Tech stack
Every proposal builds on the same PWA base, installable from the browser on any phone or computer. Dashed boxes are what this app adds.
- AppNext.js and TypeScript
- PWAManifest and service worker
- UITailwind CSS
- Data and authSupabase Postgres
- PaymentsStripe
- EmailResend
- HostingVercel
- This app · LLM providerHosted LLM API with tool useAgents need a model that can plan, call tools and write structured block outputs.
- This app · Job queueDurable background job runnerAgents run for hours and in parallel, beyond serverless request limits.
RisksAI inference
- CommercialLarge chat products could add branching or canvas views, and users may stay where they already work.
- AIAgents can produce wrong or inconsistent block outputs, and a clear audit trail may not be enough for users to trust long unattended runs.
- CommercialHours-long parallel agent runs may cost more in model usage than a $29 plan covers.
- TechnicalDependency tracking for partial reruns is hard when block outputs are free text, so stale downstream blocks may go unnoticed.
- CommercialNew users wanted chat instead of learning an interface, so the chat-to-canvas draft must work well at first use or they may leave.
Everything marked AI proposal or AI inference was written by a model from the evidence on the opportunity page. The proposed name must be checked against trademark registers before it is used.