SaaS Maker / Ideas
Scored product ideas.
A working decision ledger for tech-heavy, solo-feasible paths toward $5K MRR. Compare the wedge and the tradeoffs—not just the title.
- Ideas
- 48
- Best bets
- 32
- Dataset rule
- Technical ≥ 7
Decision workbench
Find the arguments worth testing.
How scoring works
Fun = F + T: how enjoyable and technically interesting the build is.
Money = M + Feas − C: raw potential plus solo feasibility, less competition.
A best bet has Fun of at least 14 or Money of at least 5, and a customer other than an individual developer.
| ★ | 9 | 14 | Construction daily report automation.Construction daily report automation. Foremen take phone photos + voice notes; AI generates compliance-ready daily logs (weather, headcount, work completed, safety incidents) as branded PDFs. Comp: Procore, Buildertrend are heavy ERPs; lightweight AI-first reporting is open. Wedge: target the 1–10-employee general contractor who can't justify Procore's $7K/yr. $100/site/mo × 50 sites = $5K MRR. Tech: vision-LM on photos + voice-to-text + template engine. Score breakdownF7 M7 T7 C4 Feas6 Fresh idea — feasibility hand-set on add. Fresh idea — feasibility hand-set on add. | fresh Non-developer |
| ★ | 8 | 11 | Memorial / obituary site generator for funeral homes.Memorial / obituary site generator for funeral homes. Funeral homes upload photos + bio; get a branded memorial page with comments, donation links, livestream embed. White-label per home. Comp: Tributes.com (Legacy.com — dominant but ugly), Ever Loved. Wedge: clean modern UX + AI-written tribute drafts. $50/mo per home × 100 homes = $5K MRR. Tech: white-label SaaS + media handling. Score breakdownF4 M7 T7 C4 Feas5 Fresh idea — feasibility hand-set on add. Fresh idea — feasibility hand-set on add. | fresh Non-developer |
| ★ | 7 | 16 | AI hyperlocal newsletter automation.AI hyperlocal newsletter automation. For citizen-journalists & local-newsletter ops: scrapes Reddit, city council minutes (PDFs!), event sites, local news; dedups across sources; drafts neighborhood/city newsletter with embedded sources. Comp: none doing the sourcing — beehiiv/Substack are publishing tools, not content engines. Wedge: the data pipeline. $50–100/mo × 50–100 ops = $5K MRR. Tech: distributed scrapers + LLM synthesis + dedup + scheduling. Score breakdownF8 M6 T8 C4 Feas5 Fresh idea — feasibility hand-set on add. Fresh idea — feasibility hand-set on add. | fresh Non-developer |
| ★ | 7 | 14 | Home inspector report automation.Home inspector report automation. Photos + voice notes → branded customer-ready inspection PDFs with defect categorization, severity, and repair-cost ranges. Comp: Spectora (~$60/mo), HomeGauge — established, but AI-first photo workflow is the wedge. Wedge: 5× faster report generation. Each inspector does 5–15 reports/wk; $79/mo × 65 inspectors = $5K MRR. Tech: vision-LM defect detection + RICS-style template engine. Score breakdownF7 M6 T7 C5 Feas6 Fresh idea — feasibility hand-set on add. Fresh idea — feasibility hand-set on add. | fresh Non-developer |
| ★ | 6 | 14 | AI personal CRM with auto-drafted check-ins.AI personal CRM with auto-drafted check-ins. "Haven't talked to X in 90 days — here's a draft message based on past context." Reads email + calendar, builds a relationship graph, surfaces who's slipping. Comp: Dex, Clay (B2B-focused), Monica. Wedge: auto-draft with past-context memory is the gap — others just remind. $10–30/mo prosumer/exec. 200 users at $25/mo = $5K MRR. Tech: email/calendar ingest + relationship graph + RAG drafting. Score breakdownF7 M6 T7 C6 Feas6 Fresh idea — feasibility hand-set on add. Fresh idea — feasibility hand-set on add. | fresh Non-developer |
| ★ | 6 | 13 | Etsy / Shopify niche listing optimizer.Etsy / Shopify niche listing optimizer. Upload a listing; get per-niche SEO scoring vs top sellers, photo enhancement, pricing recommendation, copy rewrite. Comp: eRank ($6/mo), Marmalead — keyword-only tools; AI-first generation is the gap. Wedge: per-niche tuning (jewelry vs digital products vs apparel). $30/mo × 170 sellers = $5K MRR. Tech: scraping top-seller signals + vision + LLM generation. Score breakdownF6 M7 T7 C6 Feas5 Fresh idea — feasibility hand-set on add. Fresh idea — feasibility hand-set on add. | fresh Non-developer |
| ★ | 6 | 12 | An app like pager-duty but for live ordersAn app like pager-duty but for live orders. Comp: PagerDuty, Opsgenie, BetterStack, custom Slackbots. Wedge: order-flow-specific (sub-minute SLA monitoring for fulfilment teams) — vertical SaaS, real but small TAM. Score breakdownF5 M6 T7 C5 Feas5 Vertical SaaS — small TAM but real; solo can build + sell into one industry. Vertical SaaS — small TAM but real; solo can build + sell into one industry. | README Non-developer |
| ★ | 5 | 17 | B4 — AI step-through debugger.B4 — AI step-through debugger. AI explains "why is this variable that value?" by tracing prior steps. Production traces via log replay. Comp: IDE debuggers, Cursor, Sentry session replay, Rookout (acquired by Dynatrace). Wedge: post-mortem replay debugging on production traces — real tech, genuinely novel. F/T are the draw. Score breakdownF8 M4 T9 C5 Feas6 Devtool, solo can ship; OSS-launch + dev community is a real distribution path. Devtool, solo can ship; OSS-launch + dev community is a real distribution path. | ai-ideas B2B technology |
| ★ | 5 | 16 | Voice memo → publish pipeline.Voice memo → publish pipeline. Record an idea; get published-ready blog post / newsletter / Twitter thread / LinkedIn post in your tone, with auto-pulled images/quotes. Comp: Castmagic ($23/mo), Cleft, Audiopen ($9/mo) — growing but specific format-per-output + tone-cloning is open. Wedge: style-transfer per channel + auto-image. $20/mo × 250 creators = $5K MRR. Tech: ASR + style-transfer + multi-format generation. Score breakdownF8 M6 T8 C6 Feas5 Fresh idea — feasibility hand-set on add. Fresh idea — feasibility hand-set on add. | fresh Non-developer |
| ★ | 5 | 12 | A2 — Legacy code resurrection.A2 — Legacy code resurrection. AI-assisted COBOL/old-Java/Perl-monolith → modern stack migration with test harness generation. Comp: Anthropic Claude Code direct push (SSA $1B project, IBM watsonx Code Assistant for Z), specialized consultancies, Azure Legacy-Modernization-Agents OSS. Wedge: Anthropic/IBM are eating the platform; deliverable wedge is still consulting + tooling — solo can land 1-2 6-figure contracts but not productize easily. Score breakdownF4 M8 T8 C5 Feas2 Enterprise sales, Fortune 500 access, multi-month cycles. Consulting is solo-doable but not a product. Enterprise sales, Fortune 500 access, multi-month cycles. Consulting is solo-doable but not a product. | ai-ideas B2B technology |
| ★ | 4 | 17 | Time-series geographical heatmap from many sources (Swiggy-style demand)Time-series geographical heatmap from many sources (Swiggy-style demand). Approximate future demand from history; IP-as-input. Timeseries DB or CH. Kalman/Viterbi for smoothing. Comp: Mapbox, Kepler.gl, Carto, Datawrapper, Foursquare Studio. Wedge: real-time multi-source forecasting (not just visualization); B2B sales motion hard for solo. Score breakdownF8 M5 T9 C5 Feas4 Tractable to demo; B2B analytics sales is brutal for solo. Tractable to demo; B2B analytics sales is brutal for solo. | README Non-developer |
| ★ | 4 | 15 | StorytunesStorytunes. Multi-author collab story writing with line-by-line voting; can include AI contributors. Each on-site event = a story; auto-summarised. Spec: docs/product/specs/storytunes.md. Comp: Storium, PlotVote, WriteJam, Storyfall, Mycelore, NovelAI (single-player). Wedge: AI persona contributors voted alongside humans — fun, niche, hard to scale revenue.Score breakdownF8 M4 T7 C5 Feas5 Same as the spec — solo-buildable, distribution is the open question. Same as the spec — solo-buildable, distribution is the open question. | README Non-developer |
| ★ | 4 | 15 | AI fact-checker / source surfacer for journalists.AI fact-checker / source surfacer for journalists. Paste a claim or article; get verified counter-sources with credibility scoring + key quotes. Comp: Originality.ai (plagiarism-focused), Logically (enterprise), Snopes (manual). Wedge: journalist/researcher tier at $30/mo with API access. 170 users = $5K MRR. Tech: source retrieval + cross-document NLI + credibility ranking. Score breakdownF7 M5 T8 C5 Feas4 Fresh idea — feasibility hand-set on add. Fresh idea — feasibility hand-set on add. | fresh Non-developer |
| ★ | 3 | 17 | Bizarre idea: fluid apps — apps that let users prompt and modify the app to their liking at runtimeBizarre idea: fluid apps — apps that let users prompt and modify the app to their liking at runtime. Comp: Bolt.new ($40M ARR in 6mo), v0, Lovable ($200M at $1.8B, $20M ARR in 2mo), Replit Agent — all build-time. Runtime malleability is genuinely greenfield. Wedge: runtime prompt-driven UI mutation, not one-shot generation. Score breakdownF8 M5 T9 C6 Feas4 Tech tractable; distribution against Lovable/Bolt's $200M+ funding is brutal. Tech tractable; distribution against Lovable/Bolt's $200M+ funding is brutal. | README Non-developer |
| ★ | 3 | 15 | StoryTunesStoryTunes. Multiplayer story engine with canon voting + AI co-authors. Comp: Storium (crowdfunded 2014, tiny), Storyfall (solo indie), PlotVote, Mycelore (preview), Sudowrite ($1.8M ARR after 5yrs, bootstrapped), NovelAI (subs only), AI Dungeon (-46% players Feb25→Apr26). Wattpad ($895M rev) owns social-serialized but not multiplayer. Reality: lane is open — no funded incumbent in voted-canon-multiplayer. But the cat is small (Sudowrite at $1.8M tells you ceiling). Wedge: real; monetization is the question. Score breakdownF8 M3 T7 C4 Feas4 | spec Non-developer |
| ★ | 3 | 14 | An agent that deep-dives on a person and finds out almost everythingAn agent that deep-dives on a person and finds out almost everything. Once it has full info it trades info from others and verifies from multiple sources, building a source-authenticity framework. Comp: Clay (massive funded play), Apollo, OSINT tools (Maltego, SpiderFoot), BeenVerified, PeopleDataLabs. Wedge: source-authenticity-from-trades primitive is novel; regulatory risk is real. Score breakdownF6 M5 T8 C6 Feas4 Legal/OSINT risk + B2B sales for revenue. Solo prototype OK; ship-to-revenue hard. Legal/OSINT risk + B2B sales for revenue. Solo prototype OK; ship-to-revenue hard. | README Non-developer |
| ★ | 2 | 17 | DB → sheet/notion app with git-style data merge, virtual grid for high-capacity, AI copilot, shareable subDataDB → sheet/notion app with git-style data merge, virtual grid for high-capacity, AI copilot, shareable subData. Comp: Airtable, Outerbase, Rowy, Glide, Baserow, NocoDB, Dolt (the actual git-for-data). Wedge: the git-style merge angle exists (Dolt) but a Notion-style UX on top is open. Hard tech. Score breakdownF8 M6 T9 C7 Feas3 Years of UX work to match Notion + merge correctness. B2B sales motion. Years of UX work to match Notion + merge correctness. B2B sales motion. | README B2B technology |
| ★ | 2 | 15 | Pluggable recommendation system anyone can plug in: custom fields, custom interactions, weightagesPluggable recommendation system anyone can plug in: custom fields, custom interactions, weightages. Vector-search backbone, beta mode, test env. Comp: Algolia Recommend (free tier 10K req/mo), Recombee ($99/mo, free tier 100K), AWS Personalize, Shaped.ai. Wedge: even "indie" tier is taken by Recombee's free plan — hard to differentiate on price alone. Score breakdownF7 M7 T8 C8 Feas3 SDK distribution + B2B sales + Algolia/Recombee free tiers undercut. SDK distribution + B2B sales + Algolia/Recombee free tiers undercut. | README B2B technology |
| ★ | 2 | 14 | A browser that remembers everything you have searched and lets you handle things accordinglyA browser that remembers everything you have searched and lets you handle things accordingly. Comp: Rewind (Mac screen recorder, EFF-audited), Mem.ai, Arc (sunset 2025 → Dia browser), Dia, Spotlight history. Wedge: Arc dying frees some attention; browser-only/lightweight + privacy is real but Dia/Browser Company will move here too. Score breakdownF7 M6 T7 C8 Feas4 Tractable but Rewind/Dia/Limitless funded and incumbent. Tractable but Rewind/Dia/Limitless funded and incumbent. | README Non-developer |
| ★ | 1 | 19 | An app that animates manga and comicsAn app that animates manga and comics. Ensure voice for a particular character remains the same, identify dialogue owners, assign a consistent voice/tone; fill scene gaps, add BGM, optionally extend novels by generating character designs. Training data: manga → anime conversions. Comp: Runway, Pika, Wonder Dynamics, AnimateAI, Komiko, Revid.ai. Wedge: manga-specific (voice consistency across many chapters + dialogue ownership detection). Real category but generic anime-gen is crowded. Score breakdownF9 M6 T10 C7 Feas2 Massive ML compute + video models + dataset + voice consistency at scale. Not solo-shippable. Massive ML compute + video models + dataset + voice consistency at scale. Not solo-shippable. | README Non-developer |
| ★ | 1 | 17 | Trustworthy automation infra (theme).Trustworthy automation infra (theme). Reliable agent execution in messy real-world systems: permissions, retries, idempotency, rollback, audits. Eval harnesses, provenance, real guardrails. Comp: LangSmith, LangFuse (acq'd by ClickHouse Jan 2026), Braintrust, Helicone, Galileo, Arize ($70M Series C Feb 2025) — extremely well-funded and consolidating. Wedge: only viable as a single vertical (e.g. revenue ops, back-office finance) where guardrails are mandatory; horizontal play is dead. Score breakdownF7 M9 T10 C10 Feas2 Theme, not product. Even narrow vertical needs deep enterprise relationships. Theme, not product. Even narrow vertical needs deep enterprise relationships. | ai-ideas B2B technology |
| ★ | 1 | 16 | Open world game with AI charactersOpen world game with AI characters. NPCs with persistent memory, emergent storylines. Comp: Inworld AI, Convai, Altera, Skyrim modding scene with LLM mods. Wedge: indie title scale; tech is fun but monetisation is brutal for solo. Score breakdownF8 M3 T8 C5 Feas3 Game dev complexity, marketing, content moderation, scaling AI costs. Game dev complexity, marketing, content moderation, scaling AI costs. | README Non-developer |
| ★ | 1 | 15 | A SM with personality bots that interact with all posts and each otherA SM with personality bots that interact with all posts and each other. Users create agents with their own API keys (for a price). Agents discuss and reach consensus. Could power StoryTunes-style collab. Comp: Character.ai, Hedra, Moltbook (AI-only social network launched 2025/26 with 37K agents in a week). Wedge: multi-agent dynamics with consensus + human-readable; Moltbook proved demand for the format. Score breakdownF8 M5 T7 C7 Feas3 Cold-start network effects, content moderation, AI costs at scale. Cold-start network effects, content moderation, AI costs at scale. | README Non-developer |
| ★ | 1 | 15 | An app that lets you build walkable 3D worlds; choose to design each buildingAn app that lets you build walkable 3D worlds; choose to design each building. Comp: Roblox Studio, Decentraland, Spatial, Horizon Worlds, Promethean AI. Wedge: AI-assisted building — but Roblox/Spatial own the install base. Score breakdownF7 M3 T8 C5 Feas3 Game-engine complexity vs Roblox/Spatial install base. Game-engine complexity vs Roblox/Spatial install base. | README Non-developer |
| ★ | 1 | 14 | AI social media with named personas.AI social media with named personas. Social network populated by AI influencer personas that post, comment, like, and cross-platform. Users can hire them. Comp: Character.ai (~$32M ARR, exploring sale), Hedra ($32M Series A May 2025), Botify. Wedge: social/feed dynamics (algorithmic timeline, persona-vs-persona arcs) — distribution will be the actual challenge. Score breakdownF7 M5 T7 C7 Feas3 Same as SM bots — cold-start + moderation + AI scaling. Same as SM bots — cold-start + moderation + AI scaling. | ai-ideas Non-developer |
| ★ | 1 | 14 | C2 — Screen / browser copilot.C2 — Screen / browser copilot. Always-on copilot that watches your screen and proactively offers context. Comp: Cluely ($20.3M raised, $7M ARR, a16z-backed, going viral), OpenInterpreter, Self-Operating Computer, Highlight, Rewind/Limitless 2.0, Granola moving into context. Wedge: privacy/local-LLM mode is plausible but Cluely sucked the oxygen out of this in 2025. Score breakdownF7 M5 T7 C8 Feas4 Cluely at $7M ARR with $20M funding sucked the oxygen out post-2025. Cluely at $7M ARR with $20M funding sucked the oxygen out post-2025. | ai-ideas Non-developer |
| ★ | 1 | 14 | magicformmagicform. AI-built forms with realtime/2-way comms + SDK embed. Comp: Typeform ($141M ARR, $935M val), Jotform ($145M, bootstrapped), Tally ($4M ARR, bootstrapped, aggressive free tier eating the low end), Fillout ($19/mo w/ AI built-in), Google Forms (47% share). Reality: every incumbent shipped AI form generation in 2024-25; the 'AI builds forms' wedge is gone. Pricing race-to-zero. Wedge: 2-way comms + embeddable SDK is real but niche. Score breakdownF7 M6 T7 C9 Feas4 | spec Non-developer |
| ★ | 0 | 16 | AI analytics company: auto-generates user reports from instrumentation; flow diagrams; identifies pain points and self-labels gapsAI analytics company: auto-generates user reports from instrumentation; flow diagrams; identifies pain points and self-labels gaps. Browser-extension mode for users to see how they interact. Can predict the user's next move. State-machine view of user lifecycle. Comp: PostHog (raised $75M Series E Sep 2025 at $1.4B, AI product assistant + session summaries shipped), Heap, FullStory, June, Mixpanel, Amplitude. Wedge: brutal market; only a sharp vertical (e.g. e-commerce conversion reports) has a shot. M_eff ≈ 1. Score breakdownF8 M8 T8 C10 Feas2 Brutal market (PostHog $1.4B); multi-year build; B2B sales. Brutal market (PostHog $1.4B); multi-year build; B2B sales. | README B2B technology |
| ★ | 0 | 16 | Productivity AppProductivity App. Life-OS combining habits, schedule, goals, journal, social mode. Comp: Notion ($400M ARR), Todoist (50M users), Motion ($50M ARR, $550M val, $75M raised), Sunsama, Akiflow, Habitica, Streaks, Habitify. Reality: life-OS scope graveyard — flexibility (Notion templates) consistently beats opinionated all-in-one. Spec is ~3 years of solo work. The dev will be the only user. Wedge: none defensible; build only if it's your daily driver. Score breakdownF8 M5 T8 C9 Feas4 | spec Non-developer |
| ★ | 0 | 15 | Tree Visualizer — visualise tree algorithms; let users build trees and write algo thereTree Visualizer — visualise tree algorithms; let users build trees and write algo there. Extend with an IDE for collaborative leetcode (test + submit for all). Comp: LeetCode, Coderpad, HackerRank, VisuAlgo, AlgoExpert. Wedge: real-time collab on the algorithm canvas; small TAM though, hard to monetise. Score breakdownF8 M3 T7 C8 Feas5 Small product, niche audience, OSS-ish, hard to monetize. Small product, niche audience, OSS-ish, hard to monetize. | README Non-developer |
| ★ | 0 | 15 | FamilyTreeFamilyTree. Social tree + auth + giftme + matrimonial + commerce + genetic angles. Comp: Ancestry ($1.3B rev/$10B mooted exit), MyHeritage, FamilySearch (free), Geni (200M profiles), WikiTree (free volunteer). Reality: records/DNA owned by incumbents; free/social owned by FamilySearch+WikiTree; MyFamily.com (the social-tree play) already shuttered in 2014. Spec scope is unshippable solo. Wedge: thin — maybe diaspora-Indian matrimonial-meets-tree only. Score breakdownF7 M5 T8 C9 Feas4 | spec Non-developer |
| ★ | 0 | 14 | Scalable n8n alternative — automation builder that doesn't fall over above N workflowsScalable n8n alternative — automation builder that doesn't fall over above N workflows. Comp: n8n ($180M Series C at $2.5B, NVIDIA-backed), Zapier, Make, Pipedream, Workato, Gumloop. Wedge: scaling-perf alone is thin — n8n's the open-source king, fighting them on their turf is brutal. Vertical RevOps angle only. Score breakdownF6 M7 T8 C9 Feas2 Fighting n8n at $2.5B + NVIDIA on their turf. Fighting n8n at $2.5B + NVIDIA on their turf. | README B2B technology |
| 7 | 17 | An on-disk trie of 1M sentences for fast next-word predictionAn on-disk trie of 1M sentences for fast next-word prediction. Disk-resident, child nodes per common prefix. Could find grammar mistakes (no matching path) or be a great suggester. Comp: none directly — niche. LLMs subsume the use case for most users. Wedge: the engineering, not the product. Score breakdownF8 M2 T9 C3 Feas8 Weekend project, ships as OSS. Not a business but a real artifact. | README Individual developer | |
| 6 | 16 | Build a minimalist graph library for funBuild a minimalist graph library for fun. Canvas, WebGL, SVG. Useful for: knowledge graphs, family-tree managers, agent decision diagrams, learning-tool annotations (click → make an anki card). Comp: D3, Sigma.js, react-flow (dominant for AI agent UIs), Cytoscape. Wedge: ergonomic API + opinionated layouts for specific use cases — but OSS-economy means M near zero. Score breakdownF8 M4 T8 C6 Feas8 OSS library, weekend-to-month, devs install for themselves. M near zero. | README Individual developer | |
| 6 | 16 | A website that tells you the time complexity of code you pasteA website that tells you the time complexity of code you paste. Tests for edge cases when given expected input size. Class to detect infinite loops, misc errors. Reports execution time on various inputs. Comp: Big-O calculators (academic), Codility-style runners, ChatGPT itself does this. Wedge: rigorous empirical testing not just static analysis. Score breakdownF7 M3 T9 C4 Feas7 Weekend project, niche devtool, freemium/$5-mo viable. | README Individual developer | |
| 6 | 14 | Let old machines be used as a serverLet old machines be used as a server. Next step Dukaan-style hosting. Full CI/CD, metrics, log management. Auto scale/descale. Comp: Hetzner/OVH self-host, Coolify, Caprover, Dokploy. Wedge: repurposing-old-hardware angle is novel for indie/hobby; weak monetization. Score breakdownF7 M4 T7 C4 Feas6 Niche infra OSS project; devs interested; weak revenue but solo-shippable. | README Individual developer | |
| 3 | 16 | How to make AI work with a new programming languageHow to make AI work with a new programming language. Given all syntax + a compiler/test suite, AI iteratively writes code. Comp: research-only (FAIR/DeepMind code-LM papers), no productized version. Wedge: academic angle + tooling for language designers; not a revenue story. Score breakdownF7 M3 T9 C3 Feas3 Pure research, no product path or market. | README Individual developer | |
| 3 | 13 | A1 — Synthetic user QA tester.A1 — Synthetic user QA tester. AI agent walks through user flows from a natural-language spec; reports regressions. Comp: QA Wolf (managed service), Mabl, Reflect, Octomind (raised $4.8M), Momentic, Bug0, plus Devin/Cursor-as-tester. Wedge: persona-based synthetic users that retain memory across sessions — credible technical wedge, but space is filling fast. Score breakdownF6 M7 T7 C8 Feas4 QA Wolf/Mabl/Octomind funded; solo can build but engineering-team distribution hard. | ai-ideas B2B technology | |
| 3 | 13 | Build something like wikiboardBuild something like wikiboard.org but for the entire net. Click-to-zoom on tab, parallel threads, AI summary, full-page-embed, side-comments on highlights, PDFs too. Comp: Glasp, Hypothesis, Recall, Heptabase, Scrintal, tldraw. Wedge: the spatial canvas (zoom + parallel) + AI summary — tldraw + AI is moving here. Score breakdownF6 M4 T7 C5 Feas4 | README Non-developer | |
| 2 | 14 | A chatbot arena where LLMs argue and reach consensusA chatbot arena where LLMs argue and reach consensus. Multi-agent decomposition of complex reasoning vs. single-model. Comp: Chatbot Arena, Poe, LMSYS, debate-style multi-agent papers, Anthropic/OpenAI multi-agent demos. Wedge: consensus-protocol angle for harder tasks; research-y, low revenue. Score breakdownF7 M3 T7 C6 Feas5 Research-y, OSS-friendly, low monetization but solo can ship. | README Individual developer | |
| 2 | 13 | Personal knowledge assistant.Personal knowledge assistant. Personal RAG that flags whether a new input is genuinely new vs. already known to you. Primitive that powers "is this novel?" elsewhere. Comp: Mem, Notion AI, Reflect, Obsidian + Smart Connections, Glasp. Wedge: API-first; let other apps consume the "novelty" signal — this is the only defensible angle. Score breakdownF6 M5 T7 C8 Feas5 Solo can build the primitive; standalone monetization unclear; better as a feature. | ai-ideas Non-developer | |
| 2 | 13 | Analyses GitHub repository history; explains how the codebase works and behavesAnalyses GitHub repository history; explains how the codebase works and behaves. Determines importance by change-frequency; explains each commit, effort, contribution patterns. Comp: Sourcegraph, Greptile ($25M Series A Sep 2025, Benchmark-led), CodeRabbit ($60M Series B), Graphite ($52M). Wedge: historical/archaeological focus vs. real-time review; the funded incumbents will eat this feature. Score breakdownF6 M5 T7 C7 Feas4 Devtool, solo can build; distribution against Greptile $25M/CodeRabbit $60M hard. | README Individual developer | |
| 1 | 13 | Interoperability / data plumbing (theme).Interoperability / data plumbing (theme). Moving data across SaaS tools, warehouses, CRMs; schema mapping, identity resolution, lineage, business-logic glue. Comp: Fivetran (acquired Census May 2025), Airbyte, Hightouch, Stitch, Segment, Workato — consolidating. Wedge: AI-native config for solo PMs is a thin wedge against acquirers with $B war chests. Skip unless you have a unique data source. Score breakdownF5 M9 T8 C10 Feas2 Theme. Fivetran/Census just consolidated; enterprise sales required. | ai-ideas B2B technology | |
| 1 | 12 | CCTV app customisable for special queries ("how long did the doctor sit?", "is the bin in position?")CCTV app customisable for special queries ("how long did the doctor sit?", "is the bin in position?"). Comp: Verkada (shipped AI-powered natural-language search Oct 2025: "person driving forklift" queries), Camio, Lumana, custom CV. Wedge: Verkada owns this for hardware buyers; software-only/BYO-camera angle is the only opening. Score breakdownF5 M5 T7 C7 Feas3 Verkada owns hardware buyers; software-only B2B sales is brutal. | README Non-developer | |
| 1 | 11 | A3 — Deep-dive investment analyst.A3 — Deep-dive investment analyst. Ingest 10-Ks, transcripts, news, expert calls; produce structured company deep-dives. Comp: AlphaSense ($4B val, acquired Tegus for $930M), Hebbia ($130M Series B), Daloopa ($13M strategic 2025), BamSEC, Brightwave. Wedge: solo investor / family office tier at $500/mo is plausible but data licensing (transcripts, expert calls) is the actual moat — and incumbents have it. Score breakdownF4 M7 T7 C9 Feas3 AlphaSense $4B with data licensing moat; solo can't get the data. | ai-ideas Non-developer | |
| 0 | 12 | B1 — AI log / incident copilot.B1 — AI log / incident copilot. NL queries on logs, anomaly explanations, cross-service correlation. Comp: Datadog Bits AI (NL log queries shipped), New Relic AI, Honeycomb Query Assistant, Splunk AI, Logz.io AI, Grafana. Wedge: vendor-agnostic ingest is interesting in theory — but log volume costs make standalone economics brutal. Score breakdownF5 M6 T7 C9 Feas3 Datadog Bits AI shipped, every incumbent racing; log-volume costs kill solo. | ai-ideas B2B technology | |
| -1 | 13 | C1 — LLM experiment / routing hub.C1 — LLM experiment / routing hub. Best-model-per-call routing, A/B, fallback chains, cost tracking. Comp: OpenRouter ($500M val, $50M ARR, 2M+ users), LiteLLM (470k downloads, Netflix/Lemonade prod), Portkey (400B+ tokens/day, 200+ enterprises), Helicone, Braintrust. Wedge: thin — OpenRouter owns the marketplace, LiteLLM owns OSS proxy. Score breakdownF6 M4 T7 C8 Feas3 OpenRouter $500M val owns marketplace; LiteLLM owns OSS proxy. | ai-ideas B2B technology | |
| -2 | 13 | Lightweight observability + logs framework you can attach to all your products with one clickLightweight observability + logs framework you can attach to all your products with one click. Comp: Datadog ($1K+ integrations), Sentry, Axiom (serverless logs darling), Highlight, BetterStack, SigNoz, Baselime. Wedge: indie-priced + dead-simple — but Axiom/BetterStack already own this space at the cheap end. M_eff ≈ 0. Score breakdownF6 M5 T7 C9 Feas2 Datadog/Sentry/Axiom/BetterStack — brutally consolidated. | README B2B technology |
Reset the controls to return to the complete ledger.