# rustman
> Solo founder building products with AI. Principles, methodology, lessons learned.

## Wiki (Knowledge Base)
- [Actix -- Rust actor framework](/wiki/actix-actor-framework): Actor model for Rust: typed messages, async/sync actors, supervision. Foundation behind actix-web. 9.2k stars.
- [ASI-Evolve: autonomous AI research agents](/wiki/asi-evolve-ai-research-agents): Agent framework for full-cycle AI research: architecture search, data curation, RL algorithms. Found 105 new attention architectures, +18 MMLU points from data strategies.
- [awesome-rust -- curated Rust library catalog](/wiki/awesome-rust): 57k-star community-maintained catalog of Rust libraries, tools, and applications. 500+ projects across 65+ categories.
- [dano -- media stream checksums via ffmpeg](/wiki/dano-media-checksums): Rust CLI that checksums internal media streams, not file containers. Stable across renames, metadata changes, container swaps. Duplicate detection for video libraries.
- [deepwiki-rs (Litho) -- AI documentation generator](/wiki/deepwiki-rs): Auto-generate C4 architecture docs from any codebase. Rust + LLM, Mermaid diagrams, multi-language, CI/CD integration.
- [Glicol -- graph-oriented live coding music in Rust](/wiki/glicol-live-coding): Live coding music language and audio engine in Rust. Runs in browser via WASM, VST plugins, embedded boards. Graph-oriented synthesis with zero-install web playground.
- [RustFS -- S3-compatible object storage in Rust](/wiki/rustfs-object-storage): Distributed object storage, 2.3x faster than MinIO for small objects. S3 + Swift API, single-node and cluster modes, Apache-2.0.
- [Stanford AI Index 2026: key numbers](/wiki/stanford-ai-index-2026): Annual AI report: $285.9B US investment, SWE-bench 60→100%, agents 12→66%, 362 incidents (+55%), talent drain -89%, 53% GenAI adoption in 3 years.
- [ERC3 — Enterprise RAG Challenge 3: AI Agents in Action](/wiki/erc3-agents-competition): Agent competition results: simple ReAct + composite tools beats multi-agent. 1st place 100% (Claude Opus, $600), 2nd 62% (GPT o3s-120B on Cerebras, $100). Evolution system, pre-execution validator, wiki rule extraction.
- [Agent benchmarks — how to measure if your coding agent actually works](/wiki/agent-benchmarks): Catalog of coding agent benchmarks: SWE-bench, PinchBench, PAC-1, LongMemEval. What they test, how scoring works, where we participate.
- [Agent patterns from practice — skills vs MCPs, dream mode, self-evolution, teaching→skill](/wiki/agent-patterns-stream2): 4 actionable patterns: skills lazy-load vs MCPs eager-load context, autonomous memory consolidation, fixed/mutable partitioning for self-evolution, interactive teaching to skill extraction.
- [Agent toolkit landscape — from prompt libraries to autonomous project managers](/wiki/agent-toolkit-landscape): Categorized comparison of AI agent toolkits: prompt libraries, executable skill systems, project managers, agent frameworks, memory systems. Growing catalog.
- [LLM providers & startup credits — where to get models, free tiers, and funding](/wiki/llm-providers-credits): Catalog of LLM providers with free tiers, startup credit programs, and open-source models. From $0 to scale.
- [MemPalace — spatial memory architecture for AI agents](/wiki/mempalace-agent-memory): Open-source local-first memory system: 96.6% on LongMemEval, palace metaphor (wings/rooms/halls), ChromaDB + SQLite knowledge graph, 19 MCP tools.
- [Token-efficient web requests — HTTP Accept header trick for agents](/wiki/token-efficient-web-requests): Reduce agent web request tokens by 80% using content negotiation. Accept: text/markdown header, html2text fallback, truncation strategy.
- [Television (tv) — fuzzy terminal search for wiki and code](/wiki/tv-terminal-search): How to set up tv (television) as interactive fuzzy finder for wiki, docs, stacks. Custom channels, ripgrep full-text search, bat preview.
- [Enterprise RAG Challenge — what actually wins in production RAG](/wiki/enterprise-rag-challenge): Enterprise RAG Challenge results: structured outputs beat vector search, top agents use router+SO+CoT patterns, no embeddings needed for top-3.
- [Tool calling internals — four layers from HTTP to parser](/wiki/tool-calling-four-layers): How tool calling actually works under the hood: HTTP API → Jinja2 chat template → xgrammar constrained decoding → Hermes XML parser. Four layers of abstraction.
- [Agent mistake → fix the harness, not the prompt](/wiki/agent-mistake-fix-harness): The single most important principle from harness engineering: when an AI agent makes a mistake, don't fix the prompt or try a better model. Fix the environment.
- [Agent self-discipline — drift detector and complexity thresholds](/wiki/agent-self-discipline): Principles for fighting AI agent degradation. The agent tends to drift into comfortable but unproductive patterns. Two mechanisms fight this:
- [Apple Foundation Models — zero-cost on-device AI for iOS portfolio](/wiki/apple-on-device-ai): Unified on-device AI layer for all iOS products (KissMyTask, FaceAlarm, VoiceJournal, QuestSchool). Zero API cost, privacy by architecture, SGR natively via ...
- [CLI-first testing — every project gets a CLI mirror](/wiki/cli-first-testing): Every project should have a CLI utility that mirrors the core business logic without UI. Not a separate tool — the same modules, different entry point.
- [Context engineering — context as code](/wiki/context-engineering): One of three harness engineering components. The idea: agent context is not chat history — it's a continuously improved knowledge base in the repository.
- [Context Graphs — agent trajectories and organizational memory](/wiki/context-graphs-summary): Context Graph — a graph capturing AI agent decision-making trajectories: not just *what* the agent did, but *how* and *why*. "A system of record for decision...
- [Decision Making](/wiki/decision-making): How to make decisions as a solo founder. From mental models to practical frameworks to psychological traps.
- [Decision traces compound — each agent action improves future ones](/wiki/decision-traces-compound): The economic argument for context graphs: agent decision traces are a compounding asset. Each successful action recorded in the graph improves future ones th...
- [Universal Development Principles — every project, any stack](/wiki/dev-principles-summary): 993-line reference document applied to every project regardless of stack. Injected into PRDs during generation. Covers: SOLID, DRY (but Rule of Three), KISS,...
- [Harness Engineering — development in the age of agents](/wiki/harness-engineering-summary): Harness Engineering — designing environments where AI agents do reliable work. Humans steer, agents execute. Synthesized from three sources: OpenAI experimen...
- [Infrastructure: two tools, serverless by default](/wiki/infra-two-tools): Unified infrastructure for multiple mini-startups. Two principles: one project = one deploy config (infra lives in the repo), two tools, no more (SST + Pulumi).
- [KMP sync — Supabase + PowerSync for offline-first cross-platform](/wiki/kmp-sync-strategy): For family/cross-platform apps needing offline-first sync between Android + iOS: Supabase + PowerSync is the recommended stack. MongoDB Realm Sync was deprec...
- [Origin story — Life2Film to SuperDuperAI, second mountain](/wiki/origin-story): Started Life2Film in 2014: help people turn everyday moments into meaningful stories. Not just videos — a way to reflect on life. Technology wasn't ready, pe...
- [PostHog analytics: one product = one project, cross-platform identity](/wiki/posthog-cross-platform): Unified analytics for solo founder with mobile apps + web landings. Core problem: one person on landing and in app = 2 anonymous IDs. Broken attribution, bro...
- [Privacy-First Architecture](/wiki/privacy-first): Building products where user data stays on-device. Offline-first, local processing, minimal cloud dependency.
- [SGR — schemas first, logic second, UI last](/wiki/schema-guided-reasoning): Schema-Guided Reasoning (SGR): all projects built from schemas to logic, not the other way around. Schemas are the contract between AI agent, business logic,...
- [Solo Methodology](/wiki/solo-methodology): AI-driven solo development — how one person ships products using LLM agents, harness engineering, and automation.
- [Startup Validation](/wiki/startup-validation): How to validate startup ideas before investing time in building. From quick signal checks to full STREAM analysis.
- [Background jobs — cron → CF Workers → Prefect → Temporal → Trigger.dev](/wiki/background-jobs-ladder): 5 tools for background jobs, ordered by complexity. Rule: start with cron, move up only when you need to.
- [Agent sandboxing — safe LLM agent execution on Mac](/wiki/agent-sandboxing): LLM agents (Claude Code, Codex, Gemini CLI) run with full filesystem access. Sandboxing via lightweight VMs solves this: agent works in isolation, sees only ...
- [Inspiration sources — curated channels, communities, newsletters](/wiki/inspiration-sources): Curated list of ~60 channels/communities across 7 platforms, organized by relevance to solo founder workflow.
- [Kill / Iterate / Scale — weekly decision rules](/wiki/kill-iterate-scale): Hard numeric thresholds for weekly product decisions. No "let's wait and see":
- [Launch Playbook — From Niche to Revenue](/wiki/launch-playbook-summary): Operational playbook in 11 sections. Core: idea → live product with traffic in 5-9 days.
- [Solo Founder Manifesto — Core Values](/wiki/manifest-summary): The manifesto (86 lines) establishes 6 foundational positions:
- [One pain → one feature → launch](/wiki/one-pain-one-feature-launch): The antidote to "platform thinking": don't build a platform, don't draw a 50-feature roadmap. One pain, one solution, shipped in days.
- [Portfolio approach — multiple small bets, not one big bet](/wiki/portfolio-approach): Build multiple small products instead of going all-in on one. Pieter Levels: 70+ projects, 3-4 active revenue sources. John Rush: 20+ products. Danny Postma:...
- [Privacy is architecture, not a feature](/wiki/privacy-as-architecture): Key distinction from the manifesto: privacy is not a settings checkbox but an architectural decision made before the first line of code.
- [RAG patterns — 7 approaches from vector to agentic](/wiki/rag-patterns): Reference catalog of 7 RAG patterns, each with when-to-use, trade-offs, and components.
- [S.E.E.D. — niche scoring in 10 minutes](/wiki/seed-niche-scoring): Quick filter for niche ideas before writing code. 4 dimensions:
- [Solopreneur success patterns — benchmarks, pricing, distribution](/wiki/solopreneur-patterns-summary): Data-driven patterns from successful indie hackers (2023-2026). Used by `/validate` for STREAM scoring and `/research` for market context.
- [CodeGraph — code intelligence across all projects via FalkorDB + tree-sitter](/wiki/codegraph-guide): Knowledge graph of source code across all projects. Scans files, symbols (functions, classes, protocols), dependencies, imports, function calls, class inheri...
- [2026 validated app ideas — homeschool, family productivity, privacy-first](/wiki/validated-apps-2026): Five underserved market intersections where no competitor operates. Score: 8.1/10. 50+ evidence sources. All privacy-first, offline-first, on-device AI.
- [Antifragile life design: barbell across domains](/wiki/antifragile-life-design): Taleb's antifragility applied not to investments but to life design across 5 domains: income, health, relationships, skills, location. Three levels per domain:
- [5-step decision framework: from epistemics to action](/wiki/decision-framework-5-steps): Every decision passes through 5 steps, each with its own checklist:
- [Ikigai as 8-9 daily micro-joys, not a mission](/wiki/ikigai-micro-joys): The Western interpretation of ikigai (intersection of "what I love / what I'm good at / what pays / what the world needs") is a distortion of the original. A...
- [Optimizer syndrome: best choice ≠ more happiness](/wiki/optimizer-syndrome): Barry Schwartz's research ("Paradox of Choice"): maximizers (those who seek the best option) objectively find better outcomes but experience less happiness a...
- [Summary: STREAM Framework — 6 Layers Decision System](/wiki/stream-framework-summary): STREAM = Scope (epistemic) · Time (temporal) · Route (action) · Exposure (stakes) · Audience (social) · Meaning (meta). Six layers arranged hierarchically — ...
- [STREAM: principles as hierarchy, not a flat list](/wiki/stream-six-layers): The key difference between STREAM and typical principle collections: it's not a flat list but a 6-layer hierarchy where each layer builds on the previous one.
- [SGR complete guide — code examples, patterns, structured output](/wiki/sgr-deep-dive): 1855-line reference compiling Rinat Abdullin's Schema-Guided Reasoning (SGR) series. Core: use Pydantic/Zod schemas as the interface between LLM and code — t...

## Projects
- [Visa Photo — biometric photo tool](/projects/project-visa-photo): Free offline desktop app for biometric visa photos. AI background removal, country presets, ICAO guides, print layout. Rust + Dioxus.
- [SuperStoryboard — AI storyboard generation with Figma sync](/projects/project-superstoryboard): AI-powered storyboard generation with real-time Figma collaboration. 2nd place at Supabase Hackathon. Gemini + Veo 3.1 + PGMQ.
- [airq — CLI air quality checker](/projects/project-airq): Check AQI, PM2.5, PM10 for any city from terminal. Open-Meteo + Sensor.Community data merge. Rust core with WASM support.
- [rust-code — AI terminal coding agent in Rust](/projects/project-rust-code): TUI coding agent with structured agent loop, fuzzy search, tmux background tasks, skills, MCP support. Alternative to Claude Code in Rust.
- [Invoice CRM — free file-based invoice generator for freelancers](/projects/project-invoice-crm): Open-source CLI for PDF invoices, QR payments (SEPA/EMVCo/SWIFT/crypto), digital signatures, encryption. File-based, git-friendly, AI-agent compatible. No SaaS, no monthly fees.
- [openai-oxide — idiomatic Rust client for OpenAI API](/projects/project-openai-oxide): 1:1 parity with openai-python. Typed responses, streaming, function calling. First Rust crate shipped via solo-factory pipeline.
- [OpenWok — transparent food delivery platform](/projects/project-openwok): Open-source food delivery with transparent pricing, Dioxus fullstack (SSR + WASM), Rust monorepo, Cloudflare Containers deploy.
- [SEO CLI — all search engine data in one terminal](/projects/project-seo-cli): Unified CLI for SEO: Google Search Console, GA4, Cloudflare, Bing, IndexNow. Launch, audit, monitor, reindex — 18 commands, one config, all sites.
- [SuperVox — voice pipeline toolkit for Rust](/projects/project-supervox): Voice-powered productivity TUI: live call assistant with real-time subtitles, post-call analysis, agent chat with history. STT, VAD, TTS in Rust.
- [You2Idea — AI startup idea search from YouTube](/projects/project-you2idea): Search 700+ YouTube videos from 46 startup channels. AI extracts ideas, generates PRDs, maps topic clusters. Terminal-style hacker UI.
- [solograph — code intelligence MCP server](/projects/project-solograph): Code intelligence MCP server: multi-project code graph, semantic search, session history, knowledge base, web search. 15 tools for Claude Code. FalkorDB + tree-sitter + MLX.
- [solo-factory — Claude Code plugin for solopreneurs](/projects/project-solo-factory): Claude Code plugin with 30 startup skills, 3 agents, 9 stack templates, MCP code intelligence. Full pipeline: research → validate → scaffold → build → deploy → launch.
- [FaceAlarm — face yoga tracker with daily selfie alarm](/projects/project-facealarm): Face yoga tracker with daily selfie alarm. AI won't turn off until you take a selfie. Track transformation daily. iOS + Android.
- [KubizBeat — Bashkir kubyz instrument app](/projects/project-kubizbeat): Virtual jaw harp (kubyz/vargan) synthesizer for iOS. Physics-based sound engine, multiple tunings from Bashkir, Yakut, Vietnamese traditions. Cultural preservation through technology.
- [SuperDuperAI — the company behind the products](/projects/project-superduperai): SuperDuperAI is the umbrella: SuperDuper AI Teknoloji (Turkey, tech park with R&D tax benefits) + SuperDuperAI Corp (USA, Delaware). Not a single product — a portfolio vehicle for building privacy-first AI tools.
- [Life2Film — turn everyday moments into meaningful montages](/projects/project-life2film): AI platform that turns everyday moments into meaningful montages. Rust core (video-analyzer, 11 crates), mobile-first, WASM for web. The original dream since 2014, 1M+ users.

## Stacks (Templates)
- [Astro Static Site](/stacks/astro-static): Astro Static Site stack template — web | typescript. Copy and use.
- [Cloudflare Workers](/stacks/cloudflare-workers): Cloudflare Workers stack template — edge | typescript. Copy and use.
- [iOS Swift](/stacks/ios-swift): iOS Swift stack template — ios | swift6. Copy and use.
- [Android Kotlin](/stacks/kotlin-android): Android Kotlin stack template — android | kotlin. Copy and use.
- [Kotlin Multiplatform](/stacks/kotlin-multiplatform): Kotlin Multiplatform stack template — ios + android | kotlin. Copy and use.
- [Libraries — curated libs by category](/stacks/libraries): Curated library reference — data pipelines, federation, actors, event sourcing, voice, OpenAI, canvas.
- [Next.js AI Agents](/stacks/nextjs-ai-agents): Next.js AI Agents stack template — web | typescript. Copy and use.
- [Next.js CRM / Data Dashboard](/stacks/nextjs-crm): Next.js CRM / Data Dashboard stack template. Copy and use.
- [Next.js Full-Stack](/stacks/nextjs-supabase): Next.js Full-Stack stack template — web | typescript. Copy and use.
- [Python API](/stacks/python-api): Python API stack template — server | python3.13. Copy and use.
- [Python ML/AI](/stacks/python-ml): Python ML/AI stack template — local | python3.13. Copy and use.
- [Python Scraper (Browser Emulation)](/stacks/python-scraper): Python Scraper (Browser Emulation) stack template — local | server | python3.13. Copy and use.
- [Rust Native](/stacks/rust-native): Rust Native stack template — local | server | embedded | rust (edition 2021, MSRV 1.75+). Copy and use.

## Posts
- [How I spent $250+ on an AI agent competition and what I learned](/posts/pac1-competition-retrospective): Personal retrospective on BitGN PAC1 challenge: scoring 17/104 live, organizing a hub event while competing, $50 per GPT-5.4 run, and rebuilding the agent to 74/104 in two days.
- [Agent-Bit: building a PAC1 competition agent in Rust](/posts/agent-bit-pac1): How we built a CRM agent scoring 93% on BitGN PAC1 benchmark using sgr-agent framework, ONNX classifiers, pipeline state machine, and 14 reusable tools. Architecture, lessons learned, tool design principles.
- [Agent Memory Architecture: From Working Memory to Compound Learning](/posts/agent-memory-architecture): 4 memory types, 3 operational loops, 5 systems compared. From dict-based v1 to spatial palaces and knowledge graphs.
- [Launch Playbook: From Niche to Revenue](/posts/launch-playbook): 5D workflow (Discovery→Decision→Design→Delivery→Distribution), S.E.E.D. niche scoring, fake-door validation, kill/iterate/scale rules.
- [Context Graphs — Agent Trajectories and Organizational Memory](/posts/context-graphs): Infrastructure layer that turns AI agent decision traces into compounding assets through a capture-retrieve-apply loop.
- [Harness Engineering — Development in the Age of Agents](/posts/harness-engineering): Design environments where AI agents do reliable work. Three components: context engineering, architectural constraints, feedback loops.
- [STREAM Framework: 6 Layers Decision System](/posts/stream-framework): Personal decision system with 6 hierarchical layers: Scope, Time, Route, Exposure, Audience, Meaning.
- [Solo Founder Manifesto](/posts/manifest): Why we build, what we believe, and how we operate. Privacy as architecture, antifragile design, one pain one feature launch.

## Skills (Solo Factory)
- [/audit](/skills/audit): Docs need health check
- [/build](/skills/build): Track ID referenced, or plan tasks need execution
- [/community-outreach](/skills/community-outreach): Need to find relevant threads for launch
- [/content-gen](/skills/content-gen): Need social posts from PRD
- [/deploy](/skills/deploy): After /build completes and project needs hosting
- [/factory](/skills/factory): First-time installation of the startup pipeline
- [/github-outreach](/skills/github-outreach): Targeting users of competitor libraries. NOT for Reddit (/reddit) or social copy (/content-gen)
- [/humanize](/skills/humanize): Text reads like AI-generated output with em dashes and stock phrases
- [/index-youtube](/skills/index-youtube): Need searchable YouTube content
- [/init](/skills/init): First session in solo-factory. Safe to re-run
- [/knowledge](/skills/knowledge): Needing methodology reference, asking about principles, harness engineering, SGR, launch playbook, agent memory, decision frameworks, or any topic covered in the solopreneur knowledge base. Also use when the user asks or
- [/landing-gen](/skills/landing-gen): Need marketing page from PRD
- [/launch](/skills/launch): Product is built and needs distribution strategy
- [/legal](/skills/legal): Project needs compliance documents
- [/memory-audit](/skills/memory-audit): Need to optimize agent context loading
- [/metrics-track](/skills/metrics-track): Need analytics plan
- [/pipeline](/skills/pipeline): Need multi-skill automation
- [/plan](/skills/plan): Describing a feature/bug/refactor, or need task breakdown before building
- [/reddit](/skills/reddit): Need to participate in Reddit threads. NOT for thread discovery (/community-outreach) or social copy (/content-gen)
- [/research](/skills/research): Need evidence before validation
- [/retro](/skills/retro): After pipeline completes and needs quality assessment
- [/review](/skills/review): After /build or /deploy completes
- [/scaffold](/skills/scaffold): Need project from PRD + stack template
- [/seo-audit](/skills/seo-audit): Need website SEO health check
- [/setup](/skills/setup): After /scaffold before /plan to generate workflow config
- [/sgr](/skills/sgr): Need to design a schema-guided reasoning pipeline for an agent or API
- [/skill-audit](/skills/skill-audit): Evaluating skill against best practices
- [/stream](/skills/stream): Facing a high-stakes founder choice
- [/swarm](/skills/swarm): Want faster multi-angle alternative to /research
- [/validate](/skills/validate): Need idea scoring with PRD output
- [/video-promo](/skills/video-promo): Need video marketing material
- [/you2idea-extract](/skills/you2idea-extract): Mining video content for business opportunities

## About
- [About](/about): Solo founder building products with AI