Middle Mann  /  Compare  /  AI Workspaces

The AI workspace for people who live in AI all day.

Cherry Cockpit is the workspace ChatGPT and Claude would need to become at 100 conversations. Real filesystem-style folders. Three tiers of reusable content: pinned messages, per-chat clipboard and project-wide snippets. Per-message cost tracked against user, workspace, chat and model. Multi-model routing with an Agent Library that names the exact model on every response. Private file storage with presigned URLs. Built for freelancers, consultants and researchers running dozens of concurrent AI projects. It scales into teams and enterprise as the work does.

Cherry nests folders to unbounded depth. Real filesystem organization, not a flat list.
3
Cherry ships three content reuse tiers: Pins, Clipboard and Snippets.
$0.00
Hosted tools show no per-prompt cost. Cherry logs every prompt.
Compare
Why It Exists

Consumer AI tools weren't built for people who use AI 20+ hours a week.

ChatGPT, Claude, Gemini, Perplexity, You.com all shipped as consumer chat products. They work great for a few questions a week. They fall apart when AI becomes the primary work environment for a professional.

01

Flat lists collapse at 50+ conversations

ChatGPT Projects don't nest. Claude Projects are flat. Perplexity Spaces are flat single-level hubs. Cherry ships an infinite nested folder tree with drag-and-drop reorganization and full-text search across every message in every folder.

02

Zero cost visibility until the bill lands

ChatGPT Team bills $25-30/user/month whether a user sends 5 prompts or 5,000. Cherry logs a UsageEvent on every message with tokens, dollar cost and latency attributed to user, workspace, chat and model. Filter by day, week, month or billing cycle.

03

Copy-paste is the only reuse mechanism

ChatGPT Custom Instructions are per-user. Custom GPTs are OpenAI-only. Cherry ships three tiers: Pins bookmark messages in a chat, Clipboard saves passages per conversation, Snippets are project-wide reusable text with tags and one-click injection.

04

One vendor per product

ChatGPT is OpenAI. Claude is Anthropic. Gemini is Google. Different tasks favor different models. Cherry's Agent Library routes across providers with tier-gated Presets and a Compare modal that fires one prompt at multiple models side by side.

05

Your work sits on someone else's servers

Every hosted tool stores conversations, files and history on vendor infrastructure. Change the terms, sunset a model, close an account and it's gone. Cherry runs on your Postgres and your MinIO. Every chat, file and UsageEvent lives on the hardware you own, portable and exportable on your schedule.

What It Actually Ships

The feature set the copy-paste-out-of-ChatGPT crowd needed years ago.

Cherry is Next.js 14 on the frontend, Express on the backend, PostgreSQL through Prisma, Redis for caching and MinIO for private S3-compatible file storage. It streams via Server-Sent Events. The data model spans fifteen-plus entities. Everything below ships today, in the box.

Organize & Reuse
01Folders. Infinite nested tree, self-referential parents, duplicate-name check per level, circular-reference prevention, delete-only-if-empty
02Pins. Bookmark any message in a chat, scroll-to-highlight animation from the right sidebar
03Clipboard. Per-chat saved passages, drag-reorderable, copy-to-OS-clipboard, go-to-source, edit-in-place
04Snippets. Project-scoped reusable text with tags, folder scoping, source-message tracking, one-click injection to any chat
05Global search. Six parallel queries across chats, snippets, files, folders, pinned messages and clipboard items, ranked
Multi-Model Intelligence
01Agent Library. Presets with model + temperature + system prompt + tier requirement across six categories
02Model comparison. Fire the same prompt at multiple models side by side, diff the responses
03Providers. OpenAI + Anthropic adapters shipping, pluggable interface for Google, xAI, Perplexity Sonar, others
04Streaming. Server-Sent Events over HTTP with cancel-mid-stream Stop button and optimistic message rendering
05Tier gating. Per-model tier requirement with lock icon in the picker, PlanRestrictedError enforced server-side
Ops, Teams & Extensibility
01Usage dashboard. Today / 7d / 30d / billing-cycle ranges, daily cost bar chart, breakdown by user / model / chat, storage GB
02Private file storage. MinIO with local-disk fallback, presigned URLs with short TTL, filename sanitization, no bucket path exposure
03Bulk folder upload. Upload an entire OS folder tree, server replicates the identical hierarchy transactionally
04Teams. Multi-workspace with roles (Owner, Admin, Member), per-workspace BillingPlan and StorageUsage
05Extensibility. Cherry outbound API, GitHub / Google Drive connectors, Canvas + Code Interpreter apps, 18 keyboard shortcuts
Aggregators & Open-Source Workspaces

Cherry vs the multi-model tools you might have already tried.

These are the closer peers. They already route multiple providers. The difference is depth: Cherry's fifteen-plus entity data model, three-tier content reuse, per-message cost attribution, private file storage discipline, multi-workspace roles, keyboard-shortcut system and modular Apps aren't there in the peer set, even in the open-source ones.

FAQ

Frequently asked questions.

What's actually in the box?

Next.js 14 App Router frontend in TypeScript with Tailwind. Express backend in TypeScript with Pino structured logging and Zod validation. PostgreSQL via Prisma. Redis for caching. Self-hosted MinIO for S3-compatible file storage with local-disk fallback. Server-Sent Events for streaming. OpenAI and Anthropic provider adapters. Fifteen-plus entity data model covering users, workspaces, memberships, projects, folders, chats, messages, snippets, clipboards, files, models, presets, usage events, billing plans and storage usage. Deployed via Docker Compose orchestrating nginx + frontend + backend + Postgres + Redis + MinIO.

Is Cherry self-hosted or cloud?

Cherry is self-hosted end to end. Deploy it on your own hardware for internal use, run it in your cloud or spin it up as your own SaaS product. Docker Compose orchestrates nginx, the Next.js frontend, the Express backend, Postgres, Redis and MinIO. You own the code, the data and the deployment.

Which providers ship out of the box?

OpenAI and Anthropic adapters ship today. The provider abstraction is a typed interface, so adding Google, xAI, Perplexity, You.com, DeepSeek or any future provider is a single-file addition. Google models exist in the catalog for Agent filtering; the runtime adapter is on the roadmap.

How does the three-tier content reuse system work?

Pins are per-message bookmarks: mark critical messages in a chat for quick jump-back, with scroll-and-highlight animation. Clipboard holds per-chat saved passages: save an excerpt with an optional note, drag-reorder them, copy to OS clipboard or click to jump back to the source. Snippets are project-wide: save a reusable prompt or code pattern with tags and inject it into any chat in the project with one click. Promote a Clipboard item to a Snippet when a per-chat note becomes a project-wide asset.

How granular is the cost tracking?

Every message generates a UsageEvent record with tokens in, tokens out, dollar cost, latency in ms, model, preset, chat, user and workspace. The Usage dashboard filters by today, 7 days, 30 days or billing cycle and shows a daily cost bar chart, breakdowns by user / model / chat, storage growth in GB and latency stats, all powered by Recharts. Consultants can attribute AI spend per client engagement in a single query.

How much does Cherry cost?

Contact sales for pricing. Every engagement is scoped to how you plan to deploy, who your users are, which provider adapters you need lit up and what support terms fit your organization. Reach us at [email protected].

Start a Conversation

Ready to work in AI the way you actually work?

Tell us which AI tools you currently use, how many concurrent projects you juggle and where you hit the wall. We'll respond within one business day with an honest read on whether Cherry fits, including whether you should stay where you are.

[email protected]