Orchestra — One memory for all your AI

The copy-paste problem

You ask ChatGPT for code, Claude for a review, Gemini for research — then manually shuttle context between them.

Memory locked in silos

Each tool only remembers inside its own walls. Last month's breakthrough is trapped in an old chat thread.

Orchestra sits underneath

One memory layer. Every conversation, note, and file indexed together — searchable, linkable, yours.

How it works

Think of it as a shared notebook for all your AI

You keep using the tools you love. Orchestra quietly saves what matters and connects the dots — without making you switch apps or learn a new chat interface.

Diagram: ChatGPT, Claude, Cursor, and other AIs connect through Orchestra into one searchable memory

ChatGPT · Codex Claude · Claude Code Gemini Cursor · VS Code v0 · Figma · +40 more

Saves & syncs automatically

Orchestra Memory Your conversations, code, docs — on your machine or your database

Any model can read & contribute

Search across everything Open the original thread Run multi-model workflows Export & backup

1

Capture

Browser extension auto-syncs 40+ sites (ChatGPT, DeepSeek, Gemini, Claude.ai). Cursor auto-ingest hook + MCP. Claude Desktop memory_stats via extension.

2

Store

Everything lands in one database on your laptop — or Postgres you control in the cloud.

3

Search

Keywords + semantic hybrid search. Filter by AI source, date, or chats-only (hide synced code files).

4

Reuse

Insert selected or Full thread into ChatGPT, Claude, or Cursor. Pipeline and consensus on past work.

Works everywhere

One database. Three ways to connect.

Web chats and desktop coding tools don't share memory by default. Orchestra bridges them — same search, same recall, whether you're in a browser tab or inside Cursor.

Browser

Chromium extension

Auto-sync web chats, Recall related memory, and Insert selected or Full thread from the Memory panel.

Tier A — ChatGPT, Claude, Gemini, DeepSeek, NotebookLM, Perplexity, Grok, Copilot, Mistral
◎ Save · ↺ Recall · ☰ Memory on supported pages
Codex on chatgpt.com · Bolt, Lovable, Figma
Tier B — 35+ more sites via generic scrape
Popup search · Chats only filter hides repo sync noise
Full message text — not search snippets

Desktop IDE

Cursor & VS Code

Search memory, recall context, save selections — while you code.

Cursor auto-save — Agent turns ingested via global hook
Recall copies context (Ctrl+Shift+R)
Hybrid search (Ctrl+Shift+O) · per-source counts on status bar
Works in Cursor, VS Code, Windsurf today
More IDEs in testing — added free for existing customers

Agents

Built-in agent memory

Included in Orchestra — Claude Desktop, Cowork, Claude Code, and Cursor Agent use the same MCP tools. No separate Claude SKU.

Claude Desktop Chat & Cowork — orchestra-memory extension
Claude Code · Cursor Agent + MCP
memory_stats · memory_search · memory_context
Setup: install-all-surfaces.ps1 in your download
Same memory as browser + IDE — one purchase

All three surfaces write to the same Orchestra memory

🌐 Browser extension + ⌨️ IDE extension + 🔌 Agent memory → Orchestra DB → Search · Recall · Pipeline

Compatible tools

Works with the AI tools you already use

Orchestra ships with a browser add-on, an editor add-on, and agent memory. All three write to the same place — pick the surface you are on.

Browser add-on

Chrome or Edge — save chats and recall memory on the web.

ChatGPT Codex Claude Claude Code Claude Design Gemini NotebookLM Perplexity Grok Copilot DeepSeek Mistral Bolt Lovable Figma vscode.dev github.dev 40+ more sites

Editor add-on

Search memory, recall context, and save notes while you code.

Cursor VS Code Windsurf More editors coming

Agent memory

Included in every Orchestra license — Claude Desktop, Cowork, Claude Code, Cursor Agent.

Claude Desktop Claude Cowork Claude Code Cursor Agent MCP · no extra fee

Claude Desktop, Cowork, Claude Code, Cursor, browser, and IDE are all included in Orchestra Local and Cloud — no add-on for Anthropic integration. Pair with Berta Seal for Jira tickets and delivery evidence (separate product).

Claude Desktop & Cowork — included in Orchestra. Install extensions/claude-desktop/claude-desktop.mcpb from your download (same package as browser + IDE). Pick your Orchestra install folder — not a URL, not a separate Anthropic connector. Guide: docs/MEMORY-SURFACES-SETUP.md

Smart search

Find what you meant, not just what you typed

Orchestra combines classic keyword search with meaning-based search. Ask “that postcode regex from last month” even if you never used the word “regex.”

When you search, Orchestra blends two approaches:

Keywords

Exact matches

Meaning

Similar ideas

Combined

Best of both

Click to go back

Every search result shows which AI it came from and when. One click opens the original ChatGPT or Claude thread in your browser.

Team of AIs

When one model isn't enough

Orchestra can ask several models the same question, run a four-agent delivery pipeline, or launch a Swarm Review — parallel Security, Performance, Maintainability, and UX specialists with one synthesized report.

Architect plans → Coder writes → Reviewer checks → Tester validates

Swarm Review

Four specialists audit the same task in parallel — security, performance, maintainability, UX — then Orchestra synthesizes one prioritized fix list.

Skills & playbooks

Save how your team works — checklists, tone, workflows. Agents inject skills automatically in pipeline, swarm, and /api/v2/chat.

Demo: saved Skills injected into Orchestra+ chat and the agent follows your PR review playbook — Skills → `/api/v2/chat` → agent follows your rules

Deep Research

Three-step research from your memory — plan questions, analyze evidence, write a shareable report. Gaps flagged explicitly.

Consensus mode

Same question to Claude, GPT, Gemini, and others. Get an agreement score plus one synthesized answer.

Shared context

Pipeline, swarm, and chat modes pull relevant memory automatically — no manual paste.

Product tiers

Orchestra vs Orchestra+

Every license includes the full memory platform and agent workflows. Orchestra+ is the optional control plane for routing, benchmarks, and measured savings.

Orchestra — every license

Cross-tool memory (40+ sites, Cursor, Claude MCP)
Swarm Review · Skills · Cookbook · Deep Research
Pipeline · Consensus · hybrid search
Browser + IDE extensions · export/import
BYOK · your SQLite or Postgres

Orchestra+ — bundled until 31 July

Model Arena — compare models with cost & latency
Savings dashboard — USD attribution per request
/api/v2/chat routing · knowledge cache · audit replay
External MCP client — connect tool servers
Budgets · benchmarks · governance logs

View sample dashboard →

Starter agent playbooks: Download 5 starter skills (JSON) · import via Setup or POST /api/skills/import-starter

Orchestra+

The universal LLM operating layer

Orchestra+ is an optional layer on top of Orchestra memory — it does not include Claude or Cowork (those ship with the normal package). Plus reuses your existing corpus on /api/v2/chat to skip repeat provider calls and shrink prompts; savings are calculated from your usage on the cost dashboard, not guessed.

Corpus reuse → measured savings

Exact prompt cache hits return prior answers at $0. Injected memory from browser/IDE/MCP reduces tokens sent. Both show up in /plus-dashboard.html as memory reuse and compression.

Knowledge cache

Repeat questions hit plus_knowledge before any provider call. Counterfactual cost uses current model rates and your prompt + corpus size — same method as LLMCO, without the old pricing bug.

Model Arena

Same prompt to multiple models — side-by-side answers with cost and latency on every response. Pick the winner, not the hype.

Cookbook — local models

Scan your GPU and RAM. Get Ollama model picks that fit your hardware. One command to pull and run with local_first routing.

External MCP tools

Connect HTTP MCP servers you already run — Orchestra+ calls their tools while memory stays in Orchestra. Configure via ORCHESTRA_PLUS_MCP_SERVERS.

Cost & governance dashboard

Memory reuse, routing, compression, and forecasts from v2 requests. View sample dashboard → · ROI calculator →

View sample dashboard Setup guide: docs/MEMORY-SURFACES-SETUP.md · scripts/install-all-surfaces.ps1 (Cursor + Claude + browser)

See it

Web app, browser extension, and IDE — same memory

The desktop app (v0.7) wraps the same web UI with a memory sidebar on Chat. The browser extension captures chats and injects full threads into composers. Cursor and VS Code use the IDE extension plus MCP for agents.

Web app & desktop shell

Chat with memory sidebar, multi-agent pipeline, consensus, hybrid search, history, and Setup.

Orchestra hybrid search with filters and results — **Search**Keywords + semantic · filter by source and date

Orchestra four-step agent pipeline — **Pipeline**Architect → Coder → Reviewer → Tester

Orchestra Setup tab with export and import — **Setup**Extension path, folder sync, export/import bundle

Orchestra Swarm Review with four parallel specialists — **Swarm Review**Security · Performance · UX · Maintainability

Orchestra Skills playbooks — **Skills**Agent playbooks · starter pack import

Demo: Skills injected into Orchestra+ chat — **Skills → chat**Playbooks flow through `/api/v2/chat`

Orchestra Cookbook local model recommendations — **Cookbook**VRAM-aware Ollama model picks

Orchestra+ Model Arena compare — **Model Arena**Orchestra+ · cost & latency compare

Orchestra Deep Research report — **Deep Research**Memory-backed research reports

Orchestra chat with any provider — **Chat**Any provider · memory bar · Add to chat

Orchestra consensus across models — **Consensus**Ask several models the same question

Orchestra conversation history by source — **History**Browse imports by LLM and open originals

Browser extension

Capture on Claude Code, ChatGPT, and 40+ sites. Search inside the ☰ Memory panel — insert one hit or the full saved thread.

Orchestra Save Recall Memory on Claude Code — **On-page bridge**Auto-sync · Save · Recall · Memory panel

Orchestra memory panel with Insert selected and Full thread — **Memory panel**Search · pick hit · Insert selected · Full thread · Copy

Orchestra extension options and custom sites — **Options**Server URL · built-in sites · custom LLM patterns

IDE extension (Cursor · VS Code)

Search memory, recall context, save selections — same database as the browser bridge.

Orchestra commands in VS Code and Cursor — **Command palette**Search (Ctrl+Shift+O) · Recall (Ctrl+Shift+R / Ctrl+Alt+R) · Save from clipboard

Everything included

Built for people who use more than one AI

A coordination platform — not another chat window.

⚡

Orchestra+

Universal LLM OS — /api/v2/chat, provider routing & fallback, cost dashboard, knowledge cache, governance audit. Same memory DB as browser + IDE + MCP.

🌐

Browser extension

Chats, Codex, v0, Claude Code — ◎ Save, ↺ Recall, ☰ Memory with Insert selected, Full thread, and Chats only search. Tier A sites get tailored extractors.

⌨️

IDE extension

Cursor, VS Code, Windsurf — search, recall to clipboard, save selection or pasted Agent chat. Same DB as browser. Free IDE upgrades as we ship JetBrains, Zed, and more.

🔌

MCP agents

Included in Orchestra: Claude Desktop, Cowork, Claude Code, Cursor — memory_context, save, search. Orchestra+ adds v2 routing API only.

🔍

Hybrid search

Keywords + meaning-based search. Filter by Cursor, Claude Code, Codex, or any web source.

📁

Folder sync

Index local code and design docs on startup — project files become searchable memory.

🗄️

Your database

SQLite locally or Postgres you control. One DB for browser, IDE, and MCP.

📤

Export & import

Portable JSON backups. CLI export / import bundle. Move laptop → cloud.

🔑

Bring your own keys

OpenAI, Anthropic, Gemini, DeepSeek, Grok, OpenRouter, Ollama, LM Studio. No inference markup.

🖥️

Windows desktop app

v0.8.0 — one installer bundles local server, web UI, browser + IDE extensions, and Berta Seal. First launch opens the Setup wizard (skills import + Cookbook) automatically.

↩️

Inject full context

Not snippets — load the full saved message or entire thread before insert. Works in Claude, ChatGPT, Claude Code web, and via clipboard in Cursor.

Your data

Two ways to keep control

Same product — you pick where the memory file lives.

On your laptop

A single file on your machine. Nothing leaves your PC unless you export or sync folders you choose. Perfect for solo work and air-gapped setups.

Your browser & Cursor

↓ saved locally ↓

📁 orchestra.db on your disk

In the cloud (your database)

Orchestra runs on a small server. Your chats live in Postgres you own — Neon, Supabase, or your own host. We run the app; you hold the keys.

Extension & API anywhere

↓ encrypted connection ↓

Orchestra app Your Postgres

Download

Run Orchestra on your PC

The v0.7 desktop app bundles local server, web UI (with Chat memory bar), and the browser extension folder. Your memory stays in SQLite on your machine. Purchase to download — the installer is delivered on your Polar order / customer portal (not GitHub; the repo is private). Add API keys after install — same as the dev setup.

Orchestra Local — one-time purchase, launch $79

🖥️

Windows desktop — Orchestra Local

One-time · Orchestra-Setup v0.8.0 · Orchestra + Berta Seal + browser & IDE extensions

$79 launch until 31 Jul 2026 · $179 from 1 Aug

Buy & download (.exe) — $79

How it works: checkout on Polar → open Customer portal → download Orchestra-Setup-0.8.0-x64.exe (64-bit, v0.8.0). Includes Berta Seal (products/berta-seal) and a Berta Seal Setup shortcut with the Seal icon. Not on GitHub — delivered on your order page only.

🌐

Browser extension

◎ Save · ↺ Recall · ☰ Memory — search, pick a hit, Insert selected or Full thread into the composer.

Included with Orchestra Local — load unpacked from extensions/browser in Chrome or Edge after install.

Get Orchestra

⌨️

IDE + agents

Cursor / VS Code extension + MCP for Agent mode. Berta Seal optional — Jira open/done on every delivery.

Orchestra + Seal bundle →

Pricing

Simple plans. You own the memory.

Pay for Orchestra — not inflated inference. LLM bills go directly to the providers you already use.

All prices in USD. Checkout on Polar is USD. Launch until 31 July 2026; standard list price from 1 August 2026.

Free upgrades: more IDEs and agent integrations are in testing — shipped to existing customers at no extra charge.

Launch

Local

Standard $179

$79

one-time · launch until 31 Jul 2026

Launch until 31 July 2026

SQLite database on your PC
Windows desktop app (v0.7) + browser extension scaffold
Browser + IDE extensions & MCP
Claude Desktop, Cowork & Claude Code — included
Hybrid search, pipeline, consensus
Orchestra+ at launch — v2 routing, cost dashboard, audit
Export / import backups
Self-host, no subscription
New IDE integrations — free upgrades

Buy Orchestra Local — $79

+ your own API keys (pay providers directly)

Popular

Cloud · your database

Standard $29/mo

$15

per month · launch until 31 Jul 2026

Project Assistant included

Hosted Orchestra app (API + web UI)
Orchestra+ included — routing, cost dashboard
You connect your own Postgres
We never store your chat data
HTTPS for extension, IDE & remote MCP
License validation included

Buy Cloud — $15/mo

+ Postgres (~$7–25/mo at Neon, Supabase, etc.) + API keys

Launch

Cloud + database

Standard $39/mo

$24

per month · launch until 31 Jul 2026

Project Assistant included

Everything in Cloud plan
Orchestra+ included
Managed Postgres included
One bill, we handle infra
Same BYOK model for LLMs
Priority onboarding support

Buy Cloud + DB — $24/mo

Infrastructure pass-through included in price

Add-on

Orchestra+ upgrade

Standard $79

$39

one-time · existing Local owners

Included free with new purchases until 31 Jul 2026

/api/v2/chat — multi-provider routing
Provider fallback & knowledge cache
Cost savings + forecast dashboard
Governance audit & request replay
Benchmarks & plugin host
For pre-0.7 Local licenses only

Buy Orchestra+ — $39

New buyers: Plus is bundled — buy Local or Cloud above · Sample dashboard

Compare

Orchestra vs. the usual workflow

Most tools are great at talking — Orchestra is built for remembering and connecting across tools.

	Orchestra	ChatGPT / Claude alone	Built-in AI memory	Mem0 / memory APIs	Notion AI
Memory across multiple AIs	Yes — all in one place	No — one product only	Per vendor, not shared	Dev-focused, not end-user UI	Notes only, not chats
Capture browser chats automatically	40+ sites + extension	Inside that app only	Limited	Via API integration	No
Codex, v0, Figma, code builders	Browser extension	Per product	No unified memory	Build yourself	Notes only
Cursor / VS Code integration	IDE extension + MCP	Separate products	Varies	Dev API only	No
Claude Code CLI memory	MCP tools	No	Claude-only	Custom wiring	No
Search all past AI work	Keywords + meaning	Within one account	Within one account	Search API	Workspace search
Insert past context into chat	Full message or full thread	Manual copy-paste	Per app	API only	No
Agent DevOps (Jira + evidence)	Berta Seal bundle	No	No	No	No
Ask multiple models one question	Consensus mode	Switch apps manually	No	No	No
Multi-step AI workflow	Architect → Tester pipeline	Manual	No	Build yourself	No
You own the database	Local or your Postgres	Vendor cloud	Vendor cloud	Depends on setup	Notion cloud
Bring your own API keys	No inference markup	Pay vendor only	Pay vendor only	Plus their fees	Bundled subscription
Export full memory backup	JSON bundle	Limited exports	No	API-dependent	Export notes
Open thread in original AI	One click from search	Native	Native	No	No
Parallel specialist review (Swarm)	Security · Performance · UX · Maintainability	No	No	No	No
Agent skills / playbooks	Versioned, searchable skills	No	No	No	No
External MCP tool servers	Orchestra+ MCP client	No	Build yourself	No	No
Measured savings dashboard (USD)	Orchestra+ dashboard	No	No	No	No
Local model hardware guide (Cookbook)	VRAM-aware Ollama picks	No	No	No	No
Side-by-side model compare + cost	Model Arena	No	No	No	No
Typical cost (beyond LLM usage)	$79 launch^* or $15–24/mo launch^*	$20–200/mo subscriptions	Included in sub	Dev tiers + hosting	$10–20/mo per seat

^* Launch pricing until 31 July 2026. Standard from 1 August 2026: Local $179; Cloud $29/mo; Cloud + database $39/mo. See also Berta Seal — standalone or Orchestra + Seal bundle ($149 launch).

Orchestra vs. all-in-one AI workspaces

Tools like Odysseus ask you to move into one app. Orchestra sits underneath the AIs you already use — and adds measured savings plus agent DevOps.

	Orchestra	All-in-one workspace (e.g. Odysseus)
Keep using ChatGPT, Claude, Cursor natively	Yes — memory bridge	Switch to one app
Cross-tool memory	40+ sites + IDE + MCP	Inside the workspace only
Measured cost savings	Orchestra+ dashboard	No
Agent DevOps (Jira + evidence)	Berta Seal	No
Email / calendar / docs built-in	Integrate via MCP & sync	Built-in

Questions

Common questions

Does the browser add-on work inside Cursor or Claude Code desktop?

Web chats use the browser add-on (Chrome or Edge). Desktop coding uses the editor add-on (Cursor, VS Code, Windsurf) or agent memory (Claude Code, Claude Desktop). All three share the same Orchestra memory.

Is Claude Desktop or Cowork a separate product?

No. Claude Desktop (Chat and Cowork), Claude Code, Cursor Agent, the browser extension, and the IDE extension are all included in the normal Orchestra package — Local, Cloud, and the Orchestra + Seal bundle. Install the orchestra-memory extension from your download; no Anthropic marketplace connector, no extra fee. Orchestra+ is only the optional v2 routing and cost layer — not Claude connectivity.

What is Orchestra+ and what does it cost?

Orchestra+ is the LLM routing layer: /api/v2/chat, provider fallback, cost dashboard, audit replay, and knowledge cache. It does not unlock Claude or Cowork — those are in base Orchestra. Plus is bundled free with Local and Cloud until 31 July 2026 (grandfathered). Early Local owners can add Plus for $39 one-time (standard $79 from 1 August 2026).

How do I connect Claude Desktop, Cursor, and web LLMs?

One guide covers every surface: docs/MEMORY-SURFACES-SETUP.md. Windows one-command install: scripts/install-all-surfaces.ps1 (Cursor hook, Claude orchestra-memory extension, browser). Claude Desktop uses a local .mcpb — pick your Orchestra folder, not a URL. Browser: load unpacked extensions/browser for ChatGPT, DeepSeek, Gemini, and 40+ sites.

How do I save Cursor Agent chat to Orchestra?

Run scripts/install-cursor-global.ps1 — a global hook auto-ingests each Agent turn to memory (no clipboard). Manual fallback: copy chat → Ctrl+Shift+P → Orchestra: Save Chat from Clipboard. Browser chats use the ◎ Save bridge on the page.

How do I install the extensions?

After installing Orchestra Local: load the browser add-on from extensions/browser in Chrome or Edge; install the editor add-on from your IDE’s extension menu (extensions/vscode). The setup guide in the app walks through each step. Reload the browser extension after updates (F5 on AI tabs).

Orchestra and Berta Seal — one product or two?

Separate products with an optional bundle. Orchestra ($79 launch) is memory + Orchestra+ across all your AIs. Berta Seal ($79/yr launch) adds Jira open/done and test evidence for agents. The bundle ($149 launch) includes Orchestra Local + Seal Solo in one checkout — best if you want memory and accountable delivery together.

Does recall insert snippets or full messages?

Full messages. In the browser, pick a hit in the ☰ Memory panel — Orchestra loads the complete saved text (character count shown), then Insert selected or Full thread into the composer. The popup’s Copy snippet also fetches full message text, not FTS snippets.

Is this replacing ChatGPT or Claude?

No. You keep using them. Orchestra remembers what they said so you — and any other model — can use it later.

Where does my data live?

On your computer by default (a local file). Cloud plans let you point to Postgres you control — Neon, Supabase, Railway, or your own server.

Do you see my chats on the cloud plan?

On the BYO-database plan, memory stays in your Postgres. We only run the app layer; your connection string is in your environment.

What about API costs?

You pay OpenAI, Anthropic, and others directly. Orchestra doesn't resell tokens or add a markup.

Will new IDE support cost extra?

No. We are testing more editors and agent integrations now. When they ship, existing customers get them free — included in your license, no upgrade fee.

Is launch pricing permanent?

No — launch pricing until 31 July 2026 is $79 (Local), $15/mo (Cloud), and $24/mo (Cloud + database). Standard list price from 1 August 2026: $179 / $29/mo / $39/mo. Buy before the deadline to lock in launch pricing.

Can I move from laptop to cloud?

Yes. Export a JSON bundle from Setup (or orchestra export), deploy Orchestra with Postgres, and import the bundle. Conversation IDs, messages, and artifacts are preserved — same memory, new home.

Agent DevOps

Add Berta Seal

Orchestra + Berta Seal bundle — memory and Jira DevOps together

Orchestra remembers what shipped. Berta Seal makes agents open and close Jira tickets with test evidence on every delivery — seal open → work → seal done, written to the same memory. Buy separately or as a bundle: launch $149 for Orchestra Local + Seal Solo (standard $279 from 1 Aug 2026).

Orchestra + Seal — $149 Seal standalone →

Ready to stop copying between AIs?

Start local in minutes. Scale to cloud when you need remote access or a team database.

Choose your plan