LIVE
IDLEAwaiting updates…
LexDev

SnipVault

Code snippet manager prototype preserved after the standalone commercial thesis was killed: a real Clerk, tRPC, semantic-search, Gist-import, billing, and VS Code extension foundation never established validated demand or enough differentiation to justify launch and ongoing support.

Competes with GitHub Gist

KilledLexDev
Key metrics

Metrics being defined

Competitive Intel

0 entries

No competitive intel entries for this product yet.

Research Hub

0 documents

No research documents for this product yet.

Roadmap

43 items
Planned23
Add `package.json` for the extension with activation events and contribution points
Add Clerk webhook handler to sync user creation/deletion with the users table
Add CSP headers and security hardening for production
Add error boundaries and loading states across all dashboard pages
Add integration tests for tRPC routers
Add keyboard shortcuts for common actions (Cmd+K search, Cmd+S save)
Add multi-file snippet support in the create/edit UI (schema supports it, UI currently assumes single file)
Add OpenGraph metadata and a landing page
Add rate limiting on API routes (especially AI endpoints)
Add snippet insertion command (paste snippet into active editor)
Add snippet sharing: public links with optional expiration
Add tree view provider for browsing collections/snippets in the sidebar
Build the billing page: current plan display, upgrade/downgrade flow, usage meters
Build the settings page: profile info, connected accounts, danger zone (delete account)
Implement collection drag-and-drop reordering (sortOrder field exists)
Implement plan-gating: enforce `PLAN_LIMITS` in tRPC procedures (snippet count cap, feature flags for AI/search/extension/import)
Implement proper `updatedAt` triggers or middleware
Publish to VS Code Marketplace
Run Drizzle migrations and verify schema against Neon (embedding vector column needs manual SQL migration)
Set up CI/CD pipeline (lint, type-check, build)
Set up esbuild/webpack bundling for the extension
SQL injection audit on raw SQL in search router (parameterize `sql.unsafe` calls)
Wire up the Gist import UI to the existing API routes
Done20
AI auto-tagging pipeline (GPT-4o-mini with retry logic)
Auth pages (sign-in, sign-up) via Clerk
Clerk auth middleware with public route exceptions
Code editor (Monaco) and code preview (Shiki) components
Collection tree, tag input, search bar/results, snippet card/grid components
Dark theme with custom scrollbar styling
Dashboard UI: stats cards, recent snippets grid, sidebar navigation
Database schema: users, workspaces, collections, snippets, snippet_files, tags, snippet_tags, device_codes
Device auth confirmation page (web side)
Drizzle ORM setup with Neon serverless driver
Embedding generation and storage (text-embedding-3-small, pgvector)
Environment validation with Zod
GitHub Gist import (paginated fetch with raw content retrieval)
Hybrid search: vector similarity (70%) + trigram text (30%) with recency/favorite boosts
Pages: dashboard, library, collections, search, billing, settings, import, snippet detail/edit/new
Similar snippets endpoint (5 nearest neighbors)
Stripe billing: plan limits, checkout session, webhook handler (checkout, subscription update/delete)
tRPC routers: snippet CRUD, collection management, tag operations, billing, hybrid search
UI components: button, card, badge, input, textarea, select, dialog, command palette
VS Code extension: device code auth, save snippet, search snippets, status bar integration

Timeline

No timeline events for this product yet.

Post-Mortem

Killed
Cause of Death

Weak Differentiation and No Validated Demand

Lessons Learned

A feature-rich developer tool is not a business without an urgent workflow, a specific buyer, and evidence that the workflow is valuable enough to displace free notes, Gists, IDE features, or team documentation. Deployment availability must also be reconciled with the canonical product before it is counted as launch evidence.

What I'd Do Differently

Validate one narrow workflow first, such as agent-context reuse or team knowledge retrieval, with real developers before building a multi-service SaaS and extension surface. Reuse the strongest code patterns elsewhere rather than reviving the generic snippet-manager thesis without named demand and a defensible wedge.

View full post-mortem →