Synapse, by enAIble
People and Agents Seamlessly Connected at Synapse
Synapse isn't a private assistant. It's a chat space where several people and an AI agent share the same conversation, the same files, and the same work. Top models at the LLM provider's own price, with no margin added and plan discounts on top. Or an open-source one for less.
100 welcome credits on sign-up · no card required
Checkout redesign
Mina (PM) · Daniel (Eng) + Synapse
The last two are Mina asking her own agent. Daniel sees neither.
Model choice
Every top model, plus open source
OpenAI sells you ChatGPT. Anthropic sells you Claude. Google sells you Gemini. Synapse puts all three in the same picker at the provider's own price, and adds an open-source model for less. Switch per room, or per message.
OpenAI
GPT-5.6 Sol · Terra · Luna
Hard reasoning through fast replies. GPT-6 Astra coming soon.
Gemini 3.1 Pro · 3.6 Flash · 3.5 Flash-Lite
Long context and multimodal
Anthropic
Claude Opus 5 · Sonnet 5 · Haiku 4.5
Long-running work and code
Open source
Qwen3.8 27B
Served and optimized on our own GPUs, at lower cost. More models to come.
Pricing
33–50% below what the LLM provider charges
We don't add a margin to token usage — a credit is the LLM provider's own published rate, converted directly. Every plan then discounts that: 33% on Lite, 50% on Pro. You pay considerably less than going to the provider yourself.
How a turn is actually billed
credits = tokens × provider_rate ÷ 1MAt 1 credit = $0.10. Input, cached input, output and reasoning tokens each carry the LLM provider's own rate.
The open-source model costs less
Qwen3.8 27B runs on our own GPUs, so it bills well below the frontier models. Credits are there for the moments a task deserves one of those.
Never metered
- Web search, crawling and browsing
- Running code and building files
- 2 GB personal and 2 GB shared storage
- Setting up a scheduled task
- Messages between people only
Speech and image generation are billed the same way, at their providers' own rates.
Plans
Free
$0
100 credits on sign-up
A full messenger, always free
Chat between people never uses credits. Welcome credits are valid for one year.
Lite
$19.99/mo
300 credits monthly
$0.067 per credit
About 33% below the LLM provider's own price.
Pro
$49.99/mo
1,000 credits monthly
$0.050 per credit
About 50% below the LLM provider's own price.
Subscriptions and credit packs are purchased in the iOS and Android apps. Credits are account-level, so they work everywhere.
Collaboration
Storage and compute are included
An agent can't do much without somewhere to keep files and run code. Synapse gives that space to the people and the agent together.
2 GB personal storage
Uploads are indexed, so the agent finds them by path or by search.
2 GB shared per room
Each group room gets its own, and it counts against nobody's personal quota.
A computer of its own
Somewhere to actually run things — pull tables out of a PDF, crunch a spreadsheet, draw a chart. Isolated, with no network access.
Everything above is shared with the room. One thing isn't.
Whisper
Ask your own agent something without leaving the room. A whisper is answered by your personal agent rather than the room's, so it reads the conversation everyone is having and your own history alongside it — and neither the question nor the answer is visible to anyone else.
Ask the basic question
When a term goes past that everyone else seems to know.
Catch up on what you missed
When you've been away a week and the room kept moving.
Keep your own backlog
When something lands on you, and only you.
Scheduling
Work that happens without you
Hand off what repeats, and let the agent pull the appointments you need to remember out of the conversation. The two are different things, so we keep them apart.
Scheduled work
Say "post a work plan at 9am on weekdays" and it's registered. Different times on different days go in as one schedule.
Your calendar
Commitments mentioned in chat get picked up automatically. Several in one message become several entries, not one.
Local depth
We go deep on each market
General models behave much the same everywhere. The difference shows up in whether the agent can reach the data a market actually runs on. Korea is where we've done that work so far — the US is next.
OpenDART
Regulatory filings and financials
KRX
Daily market pricing
Naver Maps
Places and local content
Seoul Metro
Live arrival times
Transit routing
Routes and travel times
Civic data
Parking and public facilities
The same pattern is what a new market gets.
Custom agents
We build your agent with you
This technology still isn't friendly to people who don't work in it every day. So we build it for you — the instructions, the tools it can reach, its starting files, the way the conversation opens.
Travel Planner
Plan a trip together
My Asset Coach
Review your portfolio
Career Manager
Track your career
Safety Secretary
Turn field notes into safety documents
More agents are in the works — and we build custom ones with you.
Tell us what you need
Custom agent
Tell us what it should do, who will use it and which data it reaches. We design and build it with you.
Enterprise
Rolling Synapse out across a team, with shared workspaces and one account to bill. Tell us how your organisation is set up.
Academia
Discounted access for students, educators and research groups. Tell us about your institution.
Capabilities
Included as standard
The agent doesn't stop at chat. It reads the live web, speaks, sees, and works inside the tools you already use.
Search and crawling
Real source links, fetched and read in full before they're cited. Common sites are parsed directly, which is dramatically faster.
Voice input and output
Hear answers in a natural voice, and speak instead of typing. Live call mode is coming soon.
Images
Image generation, plus reading documents and photos.
Integrations
Google Calendar, Drive, Gmail and YouTube, plus Instagram. More to come.