Curated memory files for common stacks and work styles. Pick one, sign up, and your AI tools know your environment before you type a single word.
Full-stack Next.js 15 App Router. Your AI tools know your routing conventions, component structure, ORM choice, and deployment target from message one.
stack.mdFramework, runtime, package manager, deploy targetconventions.mdFile naming, component patterns, import aliasesdb.mdORM, migration strategy, connection poolingtesting.mdVitest, Playwright, coverage philosophyMachine learning and data science context. Covers data ingestion, model training, eval pipelines, and serving infrastructure — the full stack your AI tools need to be useful.
stack.mdPython version, venv, key librariesdata.mdData sources, schemas, preprocessingtraining.mdModel architecture, training loop, eval metricsinfra.mdGPU cluster, storage, MLflow / W&B setupIdiomatic Go service context. Error handling conventions, database access patterns, gRPC vs REST decisions, and observability stack — loaded before the first question.
stack.mdGo version, module path, key dependenciespatterns.mdError wrapping, context propagation, loggingdb.mdsqlc / pgx patterns, transaction handlingops.mdDeployment, metrics, tracing conventionsCross-platform mobile context. Expo Router navigation, NativeWind styling, platform-specific patterns, and EAS Build setup — everything an AI tool needs to stop asking.
stack.mdSDK version, router setup, navigation patternsstyling.mdNativeWind, theme tokens, platform overridesnative.mdPlatform modules, permissions, device APIsci.mdEAS Build, OTA updates, TestFlight flowNode.js + Express + Postgres API context. REST conventions, middleware stack, auth strategy, and DB schema patterns fully loaded — no more explaining your architecture each session.
stack.mdNode version, bundler, key dependenciesapi.mdREST conventions, auth, rate limiting, error formatdb.mdSchema design, ORM, migration strategydeploy.mdDocker, CI/CD, environment managementContext for teams shipping AI-powered products. Prompt management, eval pipelines, cost tracking, model versioning — the operational layer most codebases leave undocumented.
models.mdWhich models, for which tasks, and whyprompts.mdSystem prompt library, versioning strategyevals.mdEval harness, metrics, golden test casescost.mdToken budgets, latency SLAs, fallback tiersContext for the one-person shop. Active projects, decisions, recurring commitments — so your AI tools act like the senior colleague who already knows everything you're juggling.
me.mdSkills, availability, rate, preferred clientsprojects.mdActive work, status, blockers, due datesdecisions.mdArchitectural and business decision logrecurring.mdRituals, reviews, periodic commitmentsAsync-first team context. Decision records, working agreements, and conventions that prevent every AI session from starting cold on who you are and how you work together.
team.mdMembers, roles, time zones, normsadrs.mdArchitecture decision record indexrituals.mdStandups, sprints, retros, on-call rotationstack.mdShared tech stack and deploy conventionsStart from scratch or modify any template after signing up. Memory files are plain Markdown — write them once, share them across your whole team.