New startup ideas · AI and software · Developer tools
startup idea
Portway
A machine-readable catalogue of every internal enterprise app, so agents can use them.
Portway is a subscription platform where a large company describes its internal applications once, in a no-code editor, as capability manifests: what each app does, which actions it exposes, who may call them, and what the fields mean.
- Software subscription
- Enterprise
- $100B+ market
- Platform others build on
- US first
3/5
venture judge
2
similar startups, last 2 years (8 all-time)
95%
of 4 nearest real companies still alive
yes
3 matching federal grants and programs
Direction supported by government programs and grants
Test it before you build it
$1,400 · 6 weeks · 20 prospects
For under $1,500 and six weeks, prove that platform engineering leads at large US insurers and retailers will pay $3,000 up front for a pilot that turns their half-broken service catalogue into manifests an off-the-shelf agent can actually use.
1Focus group: who and where
Platform engineering director or lead at a US insurer or retailer with 1,000+ employees who maintains a Backstage or ServiceNow service catalogue that is under 50% accurate, and whose leadership has asked for an agent strategy this quarter
where to find 20 · The Platform Engineering Slack community (platformengineering.org), the Backstage Discord, PlatformCon session speakers and attendees, and a LinkedIn Sales Navigator filter for 'platform engineering' director and lead titles at US insurance and retail companies with 1,000+ employees; KubeCon + CloudNativeCon North America for in-person follow-ups
2Sell first, build later
A 60-day design-partner pilot: 25 of your internal systems described as agent-readable capability manifests imported from your existing catalogue, with one agent workflow demonstrably completing tasks against 3 of them, starting on a named date
the ask · $6,000 per pilot, $3,000 invoiced up front, with a stated $2,000 per month platform subscription quoted at pilot exit
a real yes · A real yes is a signed order form plus the $3,000 invoice paid, or a signed pilot with a dated start and a purchase order in procurement. 'This should be a standard' enthusiasm, unpaid proofs of concept, and architecture review meetings do not count
3Small experiments
The first one attacks the riskiest assumption; each ends with a number that says whether to run the next.
1. Manifest walkthrough on calls
$250 · 14 days
Hand-author capability manifests for three systems every prospect recognizes (a ServiceNow instance, a Workday instance, one generic internal claims app) and build a clickable Figma of the no-code editor. Book 15 calls from Sales Navigator and the Platform Engineering Slack and walk each lead through authoring a manifest for one of their own systems live.
keep going if · 8 of 15 leads name the first system they would connect and accept a scheduled import session
2. Concierge catalogue import
$500 · 21 days
For 3 volunteer teams, take their real Backstage or ServiceNow export, hand-convert 10 systems into manifests, then demo an off-the-shelf agent (Claude with an MCP connector) completing one real task, such as opening a change ticket, using only the manifest. The team runs everything by hand behind the editor UI.
keep going if · 2 of 3 teams ask to expand past the free 10 systems and bring a second stakeholder to the readout
3. Paid pilot pre-sale
$100 · 21 days
Send a one-page order form to every prospect who saw a demo: 25 systems catalogued as manifests, one agent workflow working against 3 of them, delivered in 60 days, $6,000 pilot with $3,000 invoiced up front. Founder-led, no deck beyond the one page.
keep going if · 2 order forms signed with the up-front invoice paid
4Collect a deposit up front
Tesla took $1,000 refundable reservations for the Model 3 and $100 for the Cybertruck before building either: the deposit is the measurement, not the revenue.
$3,000
per prospect, refundable
how · A paid design-partner pilot invoiced up front: $3,000 of the $6,000 pilot collected via Stripe invoice paid by ACH, against a one-page order form signed by the platform engineering director; if procurement blocks pre-payment, accept only a signed order with a dated start and a PO number, and count it as half a yes set up: Stripe Invoicing ↗
what it reserves · One of three design-partner slots, the $6,000 pilot price locked, a named 60-day delivery date, and their two most-requested vendor systems prioritized for manifests
refund · Refunded in full if the 25 manifests and the working agent demo are not delivered within 60 days of the start date
target · 2 paid pilots (or 1 paid plus 2 signed with dated starts) from 20 conversations within 45 days
Go: build it if
2 pilots paid up front, and in the concierge imports the agent completes its task using manifests alone on at least 2 of 3 customer systems
Kill: stop if
0 paid pilots after 20 conversations and 3 free concierge imports, or every lead says the catalogue must be free and vendor-funded, which confirms the late-revenue trap with no bridge
Would you run this test?
One tap. The yes-share feeds the Demand pillar of this idea's score; nobody sees who answered.
Budgets are out-of-pocket estimates for a team of one to three, US market. Size the deposit to the deal, and check the terms before taking money in a regulated line.
Scorecard
One score that balances how trendy the idea is, the demand for it and its potential for 100x, with competition measured relative to every other idea in the catalog. Recent startup trends first, government priorities second.
52
Idea Score, 0-100 · raw 31.9 x 1.64
Warm
competition: more crowded than 32% of ideas · headwind x0.84
+1.5
government priorities, secondary (3 matching grants)
Trend
31
Is the wave forming now? 2025-26 entrants vs 2023-24, rounds since 2025, the sector's live-batch direction, the 2026 trend analyst.
- Entrants 2025-26 vs 2023-24 (similar companies)3
- Rounds announced 2025+ in the sector20
- Sector direction (live batch)50
- 2026 trend analyst50
Demand
32
Does anyone want it? YC's current RFS, companies already paid for something similar, the operator judge, founders' yes-rate in decks, readers who would run the test.
- YC asks for it (current RFS: idea / sector)30
- Someone already pays (similar companies, recent / all-time)67
- Operator judge: real pain0
100x potential
48
Can it return a fund? The venture judge (double weight), market-size and moat axes, neighbours still alive, the technologist judge.
- Venture judge50
- Market size axis100
- Moat axis20
- Neighbours still alive42
- Technologist judge25
Score = 100 x cbrt(Trend x Demand x 100x) x (1 - 0.5 x crowding) + government bonus (max 5), calibrated so the 95th-percentile idea scores 90 (order never changes). A geometric mean: a weak pillar cannot be papered over. Percentiles are among the 272 ideas in the catalog; the terms matched were machine-readable, catalogue, internal, enterprise, subscription, describes, applications.
The idea in full
- What
- Portway is a subscription platform where a large company describes its internal applications once, in a no-code editor, as capability manifests: what each app does, which actions it exposes, who may call them, and what the fields mean. Agents from any vendor read the catalogue instead of guessing at screens and undocumented endpoints, and software vendors publish manifests for their own products so enterprises inherit coverage rather than authoring it. Anyone can start free by connecting a single system; the catalogue only becomes load-bearing after several years of coverage, which is why revenue arrives late.
- Why now
- Scope (yc X26) is already selling discovery of software by AI agents, and 87 new dev tools companies appeared in 2024 and 62 in 2025, most of them agent harnesses that all need the same missing thing: a description of what the enterprise's own systems can do. Vendo (yc S26) and 21st (yc W26) point at the same gap from the product and UI side.
- Wedge: first customer and entry point
- One platform engineering team at a US insurer or retailer that already keeps a half-broken service catalogue: import it, make it agent-readable, free for the first ten systems.
- Path to 100x
- Enterprise application integration and middleware is a $100B+ market, and the platform wins if manifests become the format vendors ship by default, making every enterprise deployment cheaper for the next agent vendor. It is a platform, not a tool: the value is in what others build on the catalogue, and coverage compounds with each vendor-published manifest.
- Ceiling
- Standards without an owner get ignored, so the plausible cap is a consulting-flavoured catalogue product at $30-60M ARR built one field team at a time.
- Closest real companies, as the generator saw them
- Scope makes public software discoverable to agents; Portway does the inside-the-firewall equivalent and sells to the enterprise rather than the vendor. 21st supplies UI building blocks for agents, not a description of the enterprise's own systems.
- Main risk
- Model vendors bundle system description into their own agent platforms and the neutral catalogue has no reason to be separate.
Five judges
Each judge scores every idea in the catalog with a named rubric; the venture judge decides whether a card is shown at all (4-5 is venture-grade).
Venture investor
3/5
$100B+ integration market and vendor-published manifests are real power-law fuel, but an ownerless standard with revenue after year three is a long fuse.
Bootstrapper
1/5
Card admits revenue only after three years of catalogue coverage, with free onboarding for the first ten systems and no moat.
Operator
1/5
Value only arrives after vendors industry-wide ship manifests and coverage compounds for years, which is the whole-industry-moves-first pattern this lens refuses.
Technologist
2/5
Capability manifests are a no-code metadata format with no technical depth, and model vendors bundling system description would erase it in one release.
Risk
2/5
An unowned standard whose value depends on model vendors choosing to read it, with revenue only after three years and no moat stated.
trends
3/5
Agent-readable manifests mirror what UCP just did for commerce, but revenue-after-3-years timing collides with model vendors bundling this, the card's own stated risk.
Similar startups in the directory
Companies whose pitch matches most of the idea's terms (machine-readable, catalogue, internal, enterprise, subscription, describes, applications): 8 all-time, 2 from the last two years. Same matching as Idea Check.
AI documentation engineer – test & autoimprove stale product docs
Augmented reality platform for maintenance, operations, and training.
Open-source Retool for enterprise
The operating system for building mission-critical enterprise apps.
CRiskCo introduces AI-based on-boarding and underwriting technologies to Banks, Governments, lenders and credit providers applying credit to SMEs.
Exa Wizards is an AI startup that solves social issues through ideas and technology.
Betterdata uses AI to transform your production data into privacy-preserving and highly realistic synthetic data. Share, access and build with data safely across teams, businesses, and international borders.
Boardwalktech provides patented digital ledger technology for enterprise information management applications.
The generator's reference companies
Real companies the model named as closest when it wrote the card, with their fate. A check mark is a company the radar could verify in its directory.
Public money in this direction
US federal grants, SBIR/STTR awards and open opportunities from the radar's public-money feed, matched to the idea's terms; the sector totals give the context.
3
grants and programs matching the idea
117
startup-relevant grants in Developer tools
$67M
awarded in the sector, tracked
1
opportunities open now in the sector
National Science Foundation · BMMB-Biomech & Mechanobiology, Special Initiatives · $600K · posted 2026-08-20
- Actively Regulated Electrochemical Interfaces for Robust All-Solid-State Batteries Using Physics-Informed Learningawardmedium relevance
National Science Foundation · EchemS-Electrochemical Systems · $540K · posted 2026-07-28
National Science Foundation · Cell, Dev, & Physio · $1M · posted 2026-07-08
Market signal
What the radar sees in Developer tools: new companies by cohort year, the forming YC batch, and outcomes since the February snapshot.
Developer tools · 63 → 102 → 121 → 88 → 44 new companies 2022 → 2026 · 88% aliveYC S26: 8 in this cluster, 3% of the batch (was 5% in X26) (F26 is still forming: 24 listed)Since February, of 475 YC companies here: 5 acquired, 5 shut down, 47 rewrote their pitch
Design attributes
The card is one cell of a designed set: every axis below was chosen before the text was written, and the text had to realize it.
- Buyer
- Enterprise
- Business model
- Software subscription
- Path to 100x
- Platform others build on
- Market size
- $100B+ market
- Capital intensity
- Capital-medium (ops, field teams)
- Speed to revenue
- R&D first, revenue after 3 years
- Technical depth
- Integrations, no-code
- Go-to-market
- Self-serve
- Moat
- No moat yet
- Geography
- US first
- Regulation
- Unregulated
- Vibe
- Boring business
Listed under
An idea sits in its own sector and in any sector its text clearly touches.
More ideas like this
AI and software · Developer tools
Faultline
An incident agent that learns every production failure across its customers, then fixes yours.
Faultline runs as an on-call agent inside a large company's production environment: it reads traces, logs and deploys, reproduces the failure in a sandbox, and proposes or applies the fix.
AI and software · Developer tools
Reglet
Regulatory intelligence APIs that tell coding agents what the law requires.
Reglet is an API layer that maps code changes to regulatory obligations - HIPAA, PCI-DSS, GDPR, banking rules - in real time, so a bank's or hospital's coding agents and CI pipelines know when a diff touches regulated data flows and exactly what evidence to attach.
AI and software · Developer tools
Patchwright
An upgrade agent that keeps small software companies current, with a recipe store others fill.
Patchwright is a hosted agent that watches a small software vendor's repositories and does the maintenance work nobody staffs: dependency and framework upgrades, breaking API migrations, deprecation fixes, and the evidence trail for SOC 2, GDPR and accessibility questionnaires that their own customers demand.
AI and software · Developer tools
Appcroft
The operating layer that keeps small businesses' AI-built apps running.
Appcroft hosts, patches, monitors and integrates the AI-generated apps that small businesses now get built by freelancers and dev shops using tools like Lovable.
AI and software · Developer tools
Sovern
Air-gapped coding agent infrastructure for defense and banks; the code graph never leaves.
Sovern is an on-premise runtime and API that lets a defense program office or a bank run coding agents against its entire repository history without a single token crossing the network boundary.
AI and software · Developer tools
Miniplex
The cloud built for a billion small apps.
Miniplex is purpose-built cloud infrastructure for agent-generated software: sub-second cold-start microVMs, per-app metering in fractions of a cent, and a deploy API that coding agents and app builders call directly.
Fictional company written 2026-08-22 from MarkosWeb data; the companies, grants and numbers around it are real and tracked. Treat the idea as a research prompt, not a plan.