Born of pieces, bound by purpose.
Thought before form.
AIs from around the world, in one place. Run them with your own API keys. Have them remember across providers, arrange them in a workspace, ask them for consensus. SynquoRum is a workshop for thinking with many AIs.
Works with any AI provider
Multi-AI.
Many minds, one thread.
Send one question to many AIs at once. Have them read each other's answers. Have them debate. Ask for consensus through voting, compare them with a heatmap, talk together in a group. A place where multiple kinds of intelligence come together in a single thread.
Learn more → Multi-AIParallel 186NEW · β2.0
Run the same AI 186× in parallel — industry-leading
Run Opus 4.7 × 186, or mix Sonnet × 100 + GPT × 86 freely. Trigger with @opus*186 mention syntax.
Over 2× Antigravity 2.0's 93. Works with all major BYOK providers, natural diverse sampling for best-of-186.
Try Parallel 186 →Web IDENEW · β2.0
Full Git workflow in browser — industry-first
Monaco editor + Multi-AI chat + Prettier + Git (clone/commit/push) + .zip export. No PC needed, all in browser.
Coding experience that completes entirely in browser — not yet in Cursor or Antigravity 2.0. IDE works from mobile and tablet too.
Open Web IDE →Memory.
Write once. Every AI reads.
What you teach ChatGPT, Claude already knows. What you tell Claude, Gemini reads too. The subject of memory is not the model. It is the person using it.
Automatic injection of related memory, automatic extraction from conversation, individual editing and deletion, separation of sensitive data. Memory is something you can control.
Learn more → MemoryWorkspace.
Nodes and lines. Arrange your AIs.
Place AIs on a visual workspace. Give them roles, connect them with lines, move them as a group. Folders, notes, specifications, threads of discussion. Everything sits on the same surface, with relationships visible.
Git-like version control, a flow for proposing and approving specifications, duplication from templates. Even the work of writing specs as a team happens in the same place.
Learn more → WorkspaceAgent.
See the screen. Move the screen.
Inside a browser, on a desktop, in your own terminal. AIs that work in your place, regardless of where the work happens.
A Planner that lays out the steps. Computer Use that executes them. A CLI that runs on your own machine. A Scheduler that wakes up on its own. How much to delegate is up to you.
Learn more → AgentSkills.
Add and remove abilities.
Domain knowledge, repeated tasks, internal rules. Bundle them as a skill, attach them to an AI, call them when needed.
Build your own, take one from the library, share across the team. The same skill works regardless of which AI runs it.
Learn more → SkillsPrivacy.
Your keys. Your data.
SynquoRum runs on your own API keys. No markup on top. Mask personal information before it reaches an AI. Distribute requests across providers instead of concentrating them. Keep sensitive data in a separate vault and restore it only when needed. The level of privacy is yours to choose.
Learn more → PrivacyConnectivity.
SynquoRum, reachable from outside.
Connections to external services, an MCP server, webhooks, a public API. SynquoRum's features can be used from Claude.ai, Cursor, or other places.
Learn more → ConnectivityWhat SynquoRum offers.
AIs from around the world.
Anthropic, OpenAI, Google, Meta, DeepSeek, Mistral, xAI, and many more.
Your own API keys.
No markup from SynquoRum. Pay each AI provider directly.
32 languages.
Japanese, English, Chinese (Simplified and Traditional), Korean, French, German, Spanish, Portuguese, and others.
For individuals. For teams.
The same product, whether used alone or with an organization. Individual plans offer detailed personal settings. Team plans add member permission management, shared API keys, and audit logs.
Audit logs. GDPR. Webhooks. Public API.
The mechanisms for using SynquoRum at work are all prepared.
Minimize your AI costs.
Smart Router, Response Cache, and Prompt Compression — three layers work in the background, calling AI only when and how much you need.
Smart Router — automatically picks the best AI for your question.
Light models for simple questions, powerful models for complex reasoning. We assess difficulty and route accordingly.
Response Cache — never re-ask the same question.
Your past responses cached privately to your account. Same question twice returns instantly with zero tokens. Never shared across users.
Prompt Compression — industry SOTA via LLMLingua-2.
Microsoft's LLMLingua-2 integrated. Long prompts compressed while preserving meaning.
Quality Slider — choose your balance, 5 tiers.
Max Quality / High / Balanced / Aggressive / Max Savings. Pick the balance that fits your daily use, or push for maximum savings.
Your AI, in your place.
Four ways to run AI wherever you want — SaaS APIs, home GPUs, in your browser. All from the same screen.
BYOK — Major AIs supported with your own API key.
Anthropic, OpenAI, Google, Meta, DeepSeek, Mistral, xAI, and more. Pay directly to each provider, no SynquoRum markup.
BYOM — local LLMs on your PC.
Ollama, LM Studio, llama.cpp. Call LLMs running on your own machine directly from SynquoRum.
BYOC — production-grade on your home GPU.
vLLM, TGI, oobabooga — connect your Docker-based inference server to SynquoRum via ngrok / Cloudflare Tunnel.
BYOB — AI that runs in your browser.
Browser-native inference using WebGPU. Your data never leaves your PC — the ultimate privacy. (Rolling out gradually.)
Student discount for students and educators.
Sign up with your educational email and get up to 50% off on all paid plans. No time limit, applied automatically.
Eligibility — major educational TLDs.
.edu (US), .ac.jp (Japan), .edu.au (Australia), .ac.uk (UK), and other recognized educational TLDs. Auto-detected at OAuth sign-up.
Up to 50% off.
Standard $25 → $12.50 / Pro $50 → $25 / Maximum $150 → $75. When eligible for both Beta Founder Discount (20%) and Student Discount, the Student Discount takes precedence (more favorable).
Automatic — no coupon code needed.
The discount activates the moment you sign up with an educational email. Applied automatically at checkout, no extra steps required.
Standard features.
The standard parts of a SaaS are also in place.
- Image — generation, editing, recognition, variations, upscaling
- Voice — real-time conversation, transcription, text-to-speech, wake word
- Code — generation, execution, highlighting, multiple languages
- Search — web search, deep research, summaries with citations
- Files — PDF, Excel, images, arbitrary formats
- Scheduler — recurring tasks, cron, custom functions
- Pages — Notion-style documents, databases, graphs
- Meetings — minutes, decisions, calendar
- And more — tags, pins, prompt library, notifications, export
Pricing.
Individual plans come in four tiers: Free ($0), Standard ($25), Pro ($50), Maximum ($150). Choose by the scale of your use.
Team plans are calculated as "the individual plan's monthly price + $3 × number of members".
Example: Team Standard with 5 members = $25 + $3 × 5 = $40/month.
Privacy Lull (auto PII masking) is a free standard feature in all plans. Privacy Silence is an add-on for Standard or higher.
During the Beta period, use BETA_FOUNDER_20 for 20% off × 12 months.
See pricing page →