Quickstart: SynapseX Chat
This page gets you signed in to SynapseX Chat and to a first useful answer in about two minutes. By the end you will have asked a question about a document you uploaded — a contract, a spreadsheet, a report — and had the model answer from its actual contents.
Step 1: Open SynapseX Chat
Go to https://chat.synapsex.ai and click Login
in the top right, or Sign In at the bottom of the sidebar. Both take you to
https://chat.synapsex.ai/auth.
Step 2: Sign in with Google
On the sign-in page, choose Continue with Google. On first use this also creates your account.
The page also shows an email and password form. That form only signs in an account that already exists — there is no signup form in the chat app, so Continue with Google is the only self-service way to create an account here.
Step 3: Send your first message
You land back in the chat. In the composer, the model chip already shows
SynapseX Atlas (synapsex:synapsex-atlas-7b-v1), the default model.
Type a question and press Enter. Use Shift+Enter for a newline inside a message. A single message can be up to 100,000 characters.
Step 4: Ask a question about your own document
This is where the chat starts earning its keep.
- Click + in the composer.
- Choose Add files or photos.
- Attach a PDF or a spreadsheet — up to 10 MB per file.
- Ask a question about what is in it, for example:
Summarise the payment terms in this contract and list every deadline.
PDF, DOCX, XLSX, XLS, CSV, TXT, MD and JSON attachments are converted to text on the server and inlined into your prompt, so the model answers from the real contents of the file rather than from the file name. The first 60,000 characters of each document are used; anything past that is truncated.
Step 5: Switch models when the question gets harder
Press Cmd+Shift+P to open the model picker. Six models are available: Atlas, Forge, Vault, Nexus, Oracle and Quasar. The choice is saved on that conversation, so you can keep a cheap fast model for everyday questions and switch to a stronger one only where it matters.
Which one to pick for which job is covered in Models.
Three shortcuts worth learning immediately
| Shortcut | What it does |
|---|---|
Cmd+Shift+U | Start a new chat |
Cmd+K | Search your chats |
Cmd+B | Toggle the sidebar |
On Windows and Linux, use Ctrl in place of Cmd for search and the sidebar.
About the lightning chip (Turbo)
The lightning chip in the composer is already set to Auto. On questions it judges hard, SynapseX spends extra compute — generating several candidate answers and returning the strongest merged result — and leaves easy questions alone. One constraint worth knowing now: a Turbo turn cannot use tools, so switch Turbo Off for a turn where you want the assistant to look something up. Details are in Turbo.
Next steps
- Models — what each of the six models is good at
- Files and documents — supported types and limits
- Conversations and projects — keeping long-running work organised
- Chat plans — what each tier includes