This is a public excerpt of our internal vision document. Some sections are kept private — we like to cook with the kitchen door open, but not all the way.
Slopdrop Vision Document
_The home for AI-generated artifacts_
Last updated: 2026-02-24 Status: Foundation document Owner: Brand/Marketing/Shipping Lead
Why Slopdrop Exists
AI agents generate mountains of output: documents, logs, plans, code reviews, analysis reports. This stuff is too noisy for git, but too valuable to lose. Teams need:
- Persistent URLs that don't break
- Secure sharing without friction
- A place that's not "yet another folder"
The Problem Today
| Current Workaround | Why It Fails |
|---|---|
| Dump it in git | Pollutes history, annoying PRs |
| Google Drive | Wrong tool, no DX |
| Build your own S3 bucket | Now you're maintaining infrastructure |
| Paste in Slack | Gone in 90 days, unsearchable |
The Product
CLI-first document storage for AI-generated artifacts.
Core features:
- CLI-first:
slop drop report.md→ instant shareable link - Watch & auto-sync:
slop watchsyncs on every save - Global watcher:
slop watch --globalwatches all projects across tenants - Beautiful markdown rendering with ToC, callouts, tables, code blocks
- Multi-tenant CLI: one tool, many orgs, separate credentials
- Private by default, share via expiring links
- Full tenant isolation — your data never touches another org's
- Secret scanning on every upload (API keys, tokens, credentials)
- Stdin piping: pipe from any AI agent or tool
- Git remote detection: auto-infers project from repo
- Checksum-based sync: only uploads what changed
- Enterprise: SSO, audit logs, compliance (roadmap)
Primary users:
- Teams (5-50 devs) drowning in AI-generated docs
- Enterprises needing audit trails + access control
- Solo devs who want public sharing with persistent links
Success metric: Becomes the default drop target in agent workflows. "Stack Overflow for AI artifacts" — searchable, indexable, referenced.
Core Principles (Non-Negotiables)
These are existential. No exceptions.
| Principle | Why It Matters |
|---|---|
| Setup in 10 seconds or less | No config files, no account creation friction |
| Zero data leaks, ever | One breach = game over |
| Full tenant isolation | Your data never touches another org's data |
| Always secure, always on | Zero-trust architecture, 99.9% uptime minimum |
| Nice tools for technical people | Craft matters. The 1% details matter. We give a damn. |
The Soul of Slopdrop
Built by someone who uses it every day. Not "enterprise-grade" as a buzzword — craft. The 1% details that tell you someone actually cared.
What We're NOT
Explicit anti-goals to prevent scope creep:
- ❌ Not a general file storage service (use S3/Dropbox)
- ❌ Not a code repository (use GitHub)
- ❌ Not a wiki (use Notion/Confluence)
- ❌ Not trying to replace git for code
- ❌ Not building features before they're needed
Phase 1: slopdrop.dev MVP — Shipped
Shipped Features
| Feature | Status | Description | |
|---|---|---|---|
| CLI tool | Done | slop drop, list, cat, link, status |
|
| Watch mode | Done | slop watch — auto-sync on save |
|
| Global watcher | Done | slop watch --global — all projects at once |
|
| Multi-tenant CLI | Done | Per-tenant credentials, switch with --tenant |
|
| Authentication | Done | OAuth PKCE, token storage, multi-tenant | |
| Private drops | Done | Default state for all uploads | |
| Shareable links | Done | Expiring links with slop link |
|
| Team workspaces | Done | Full tenant isolation, TenantDb wrapper | |
| Secret scanning | Done | API keys caught on upload | |
| Markdown rendering | Done | ToC, callouts, tables, code blocks | |
| Stdin piping | Done | `cat doc.md \ | slop drop -` |
| Project detection | Done | Auto-infer from git remote | |
| Multi-source sync | Done | .slop config with multiple source dirs |
|
| Checksum optimization | Done | Only uploads changed files | |
| Rate limiting | Done | KV-based with circuit breaker | |
| HMAC API key hashing | Done | SHA256 + pepper, secure key storage | |
| File size limits | Done | 1MB free, 10MB paid, 1GB enterprise |
In Progress
| Feature | Status | Description |
|---|---|---|
| Homebrew distribution | Planned | brew install slopdrop/tap/slop |
| Stripe payments | Planned | Checkout, webhooks, customer portal |
| Audit logging | Partial | Only secret.bypass logged so far |
Explicitly Deferred
| Feature | Reason |
|---|---|
| Public search/discovery | Phase 2 after trust is established |
Architecture Principles
- Zero-trust from day one
- Hard guardrails on tenant isolation (no cross-prefix leaking, ever)
- Leverage existing tools (Cloudflare stack)
Competitive Position
Why not just use existing tools?
Anyone can build this with AI. But teams choose trusted, maintained solutions because:
- Trust — Security posture matters for team adoption
- Maintenance — Nobody wants to maintain yet another internal tool
- DX — Best-in-class CLI, zero config, just works
- Network effects — Searchable public corpus (future)
The Moat
- Developer experience that feels _nice_
- Security that's actually secure (not just claims)
- Brand that resonates with the AI-native developer
The Slopdrop Promise
"AI output is messy. We embrace it. We secure it. We make it findable."
Document Index
| Document | Purpose |
|---|---|
| [[VISION]] (this doc) | Core product vision and principles |
| [[ROADMAP]] | What we're building and when |
| [[specs/architecture]] | System architecture and security model |
| [[specs/markdown-rendering]] | Supported markdown features |
| [[specs/secret-scanning]] | Upload scanning rules |
| [[guides/getting-started]] | First-time setup |
| [[guides/cli-reference]] | All CLI commands |
_This document is the foundation. All product decisions should trace back to these principles._