New startup ideas · AI and software · Developer tools
startup idea
Mendrix
Autonomous agents that take over maintaining your legacy systems.
Mendrix builds an agent fleet that assumes responsibility for maintaining aging enterprise codebases - old Java, COBOL adjacents, brittle internal systems - fixing incidents, patching vulnerabilities and keeping dependencies current under an SLA, priced per system.
- AI agent as a service
- Enterprise
- $10-100B market
- Network effects, winner takes most
- US first
3/5
venture judge
27
similar startups, last 2 years (43 all-time)
93%
of 2 nearest real companies still alive
yes
8 matching federal grants and programs
Direction supported by government programs and grants
Test it before you build it
$1,500 · 6 weeks · 20 prospects
Proves a Fortune 1000 engineering leader will pay $5,000 now toward handing legacy maintenance to an outside agent vendor, in 6 weeks for $1,500.
Riskiest assumption · An enterprise engineering leader will pay real money now and commit to a dated pilot that hands maintenance of production legacy systems to an outside agent vendor, rather than keep paying the incumbent hourly outsourcer they already blame.
1Focus group: who and where
VP of Engineering or application-services director at a US enterprise of 5,000+ employees carrying 10+ orphaned Java-era internal services whose last maintainers left, currently paying an outsourcer by the hour, with an outsourcing renewal or a production incident on one of those systems in the past quarter.
where to find 20 · Warm introductions through the founders' seed investors and former enterprise colleagues, asking specifically for 'the person who owns the systems nobody wants' (channel); the Rands Leadership Slack, 30,000+ engineering leaders with active channels on legacy and platform pain (community); SHARE, the enterprise IBM Z user group association, via its member community and event attendee network of exactly the shops running COBOL-adjacent stacks (association and its directory).
2Sell first, build later
A $5,000 two-week legacy triage: founders assess five orphaned systems, deliver an incident-risk map and a fixed monthly price to assume maintenance of them under SLA - followed, for buyers who want it, by a $15,000 per quarter pilot with a dated start.
the ask · $5,000 fixed fee per assessment, invoiced at SOW signing; pilot at $15,000 per quarter for five systems
a real yes · A real yes is a paid $5,000 invoice or a countersigned pilot LOI with a start date within 60 days. 'This is fascinating', unpaid proof-of-concept requests and offers to 'champion it internally next year' are not yes.
3Small experiments
The first one attacks the riskiest assumption; each ends with a number that says whether to run the next.
1. Paid triage assessment pitch
$800 · 21 days
Founders book 15 calls through warm intros and the communities above and pitch one thing: a $5,000 fixed-fee, two-week assessment of the prospect's five worst orphaned systems - incident history review, dependency and vulnerability risk map, and a fixed monthly quote to take the systems over under SLA. Founders deliver it by hand; no product is required or promised yet. Includes one on-site visit for the most engaged prospect.
keep going if · 5 of 15 leaders request the assessment SOW
2. SOW-to-paid-invoice conversion
$400 · 21 days
Send each requester a one-page, lawyer-reviewed SOW with a net-15 invoice payable at signing, before work starts. This isolates whether their organization can and will move money to an unknown vendor without a full procurement cycle, which is the whole question.
keep going if · 3 of 5 SOWs signed and invoices paid within net-15 terms
3. Pilot LOI with dated start
$300 · 14 days
Deliver each paid assessment in person or on a call, and close with an LOI for a $15,000 per quarter maintenance pilot covering the five assessed systems, starting within 60 days, with the incident-resolution metric that decides renewal written in. The LOI needs a named start date to count.
keep going if · 2 of 3 assessment buyers sign the LOI with a start date
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.
$5,000
per prospect, refundable
how · An up-front invoice for the $5,000 assessment against a one-page SOW the engineering leader signs - a paid design-partner engagement rather than a product deposit, because a Fortune 1000 cannot credibly prepay for an agent fleet that does not exist, but a director can approve a $5,000 services invoice inside their own signing authority without triggering procurement. set up: Stripe Invoicing ↗
what it reserves · One of three October assessment slots and first priority on a pilot start date once the fleet takes its first systems.
refund · Refunded in full if the assessment report is not delivered within 15 business days of kickoff.
target · 3 paid $5,000 assessments from 15 pitches within 6 weeks
Go: build it if
3 paid assessments and 2 signed pilot LOIs with start dates inside 6 weeks: the buyer will move money and systems, so the 3-year R&D bet has a customer at the end of it.
Kill: stop if
0 invoices paid after 15 pitches and 5 SOWs sent, or every interested buyer converts the ask into a free proof-of-concept: enterprises will not pay an outsider to own these systems, and the idea dies regardless of model quality.
5 Scripts to run itoutreach message, landing copy, deposit terms · click to open
outreach message
You're almost certainly carrying a set of Java services nobody left on the team understands, and paying an outsourcer by the hour every time one of them breaks. I run Mendrix: for $5,000 flat we spend two weeks triaging your five worst orphaned systems and hand you an incident-risk map plus a fixed monthly price to take them over under SLA. Open to 20 minutes this week to see whether your portfolio fits the October slots?
landing page
Fixed-price maintenance for the systems nobody on your team understands. A $5,000 two-week assessment of five legacy systems, then a fixed monthly SLA quote to take them over. Book one of three October assessment slots.
deposit terms
The $5,000 assessment fee is invoiced at SOW signing, net 15, and reserves one of three October slots. You receive an incident-risk map for five systems and a fixed-fee takeover quote within 15 business days of kickoff, or the fee is refunded in full. No production access is required until you approve the read-only scope.
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.
58
Idea Score, 0-100 · raw 36.1 x 1.61
Crowded
competition: more crowded than 90% of ideas · headwind x0.55
+4.1
government priorities, secondary (55 matching grants)
Trend
48
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)99
- Rounds announced 2025+ in the sector16
- Sector direction (live batch)50
- 2026 trend analyst25
Demand
60
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)100
- Operator judge: real pain50
100x potential
70
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 axis67
- Moat axis100
- Neighbours still alive54
- Technologist judge100
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 autonomous, maintaining, legacy, builds, fleet, assumes, responsibility, aging.
The idea in full
- What
- Mendrix builds an agent fleet that assumes responsibility for maintaining aging enterprise codebases - old Java, COBOL adjacents, brittle internal systems - fixing incidents, patching vulnerabilities and keeping dependencies current under an SLA, priced per system. The first three years are R&D: program-analysis models, an evaluation harness and reinforcement on real fix outcomes until reliability clears the bar where a CIO will hand over a production system. Founders sell the first contracts personally, with forward-deployed engineers running onboarding.
- Why now
- Superlog (make software self-healing) and Superset, which raised $11.5M from 35 investors in August 2026, show self-improving maintenance is investable now, but both sell tooling; the contractual take-over-the-system model is untried while enterprises still pay outsourcers by the hour.
- Wedge: first customer and entry point
- One Fortune 1000 company with a portfolio of orphaned Java services and no team that understands them: a fixed-fee pilot maintaining five systems, instrumented to prove incident resolution beats the incumbent outsourcer.
- Path to 100x
- Enterprise application maintenance outsourcing is a tens-of-billions market currently priced by the hour, and an agent fleet re-prices it at software margins. Every incident resolved anywhere in the fleet trains the shared model, so the largest fleet has the best reliability record - the only thing buyers evaluate - and share compounds toward winner-takes-most, with the proprietary corpus of incident-to-fix trajectories as the asset no entrant can replicate.
- Ceiling
- If each enterprise's legacy stack is too idiosyncratic for cross-customer learning to transfer, the network effect dissolves and Mendrix caps out as one competent vendor among many.
- Closest real companies, as the generator saw them
- Superlog and Superset build self-healing and self-improving software tooling that engineering teams operate themselves; Planeo bridges code and runtime for copilots; none takes end-to-end contractual ownership of legacy maintenance.
- Main risk
- Autonomous fix reliability plateaus below the SLA threshold and every account quietly becomes a services engagement whose cost tracks engineers.
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
Tens of billions repriced at software margins is a real prize, but three years of R&D and forward-deployed engineers risk becoming a services firm.
Bootstrapper
1/5
Three years of R&D and forward-deployed engineers before a CIO hands over production systems is the opposite of near revenue.
Operator
3/5
Enterprises already pay outsourcers by the hour for this, but three years of R&D before a CIO hands over production delays the proof.
Technologist
5/5
Three years of program-analysis models and reinforcement on real fix outcomes builds an incident-to-fix trajectory corpus no entrant can buy.
Risk
2/5
Three years of R&D before revenue while contractually owning Fortune 1000 production systems under SLA, concentrating both customers and liability.
trends
2/5
'Agents can now fix code' was equally arguable in 2024, and three R&D-first years mean any timing window closes before revenue starts.
Similar startups in the directory
Companies whose pitch matches most of the idea's terms (autonomous, maintaining, legacy, builds, fleet, assumes, responsibility, aging): 43 all-time, 27 from the last two years. Same matching as Idea Check.
Full-stack Service Layer for Robots and Autonomous Machines
The AI Agent Insurance Carrier
AI Revenue Engine for the Equipment Industry
Autonomous Drones for Earth Observation
Robotire is automating vehicle maintenance for today and future…
Cosmic is an AI-powered headless CMS
An AI-operated logistics company.
The best cloud VMs for billions of agents
An autonomous data warehouse run entirely by AI agents that build pipelines, maintain data, and discover business opportunities.
Air Defense as a Service for Civilian Infrastructure.
AI that turns any camera to an autonomous worker with real-time vision
Autonomous construction on Earth and beyond
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.
8
grants and programs matching the idea
122
startup-relevant grants in Developer tools
$70M
awarded in the sector, tracked
1
opportunities open now in the sector
- An Al Facilitated Self-Administered Technology-Focused Functional Skills Assessment and Training Program for Older Adultsawardhigh relevance
NIH / NIA · SBIR phase II · $1M · posted 2026-09-22
- VINES: Track 2: OMNISCIENT: Enabling BVLOS for On-demand Mission Critical Aerial Networks through Self-adapting NextG IAB Meshawardhigh relevance
National Science Foundation · TIP-CHIPS KTA-6 Communications · $5M · posted 2026-09-18
- VINES: Track 1: Enabling Real-Time Agricultural Robotics with AI-Native Cellular Access Edgeawardhigh relevance
National Science Foundation · Use-Inspired NextG · $1M · posted 2026-07-30
NIH / NIA · STTR phase I · $250K · posted 2026-09-23
- RadBot: Next-Generation Drug Discovery for Environmental Stress Resistance Using Longevity Biologyawardhigh relevance
NIH / NIA · SBIR phase I · $300K · posted 2026-09-22
- Automated MRI-Based Musculoskeletal and Body Composition Assessment for Healthy Agingawardhigh relevance
NIH / NIA · SBIR phase II · $1M · posted 2026-09-18
NIH / NIDDK · SBIR phase II · $325K · posted 2026-09-15
- A Novel Immunological-Directed Live Biotherapy Product for Treating Crohn's Diseaseawardhigh relevance
NIH / NIDDK · SBIR phase II · $1M · posted 2026-09-01
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 · 64 → 105 → 121 → 90 → 57 new companies 2022 → 2026 · 87% aliveYC F26 live: 8 in this cluster, 6% of the batch (was 3% in S26)Since February, of 474 YC companies here: 6 acquired, 8 shut down, 59 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
- AI agent as a service
- Path to 100x
- Network effects, winner takes most
- Market size
- $10-100B market
- Capital intensity
- Capital-medium (ops, field teams)
- Speed to revenue
- R&D first, revenue after 3 years
- Technical depth
- Deep tech: ML, hardware, bio
- Go-to-market
- Founder-led sales
- Moat
- Data moat
- 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 2026-08-26 from MarkosWeb data; the companies, grants and numbers around it are real and tracked. Treat the idea as a research prompt, not a plan.