The Synsci CLI dashboard at cli.syntheticsciences.ai is the source of truth for everything the CLI doesn’t store locally, credentials, credits, sessions, team access, and audit logs. The CLI talks to the dashboard over an authenticated channel at session start and reports usage back as you work.

​
What the dashboard does

  • Credentials. Paste in tokens for Hugging Face, W&B, Modal, Lambda Cloud, AWS, GCP. The CLI pulls them into session memory at launch.
  • Credits and billing. See your balance, top up credits, set alerts, and view per-model and per-session spend.
  • Sessions. Every CLI session shows up with model usage, cost, and metadata. Investigate a runaway agent loop or audit a teammate’s run.
  • Team management. Invite teammates, assign roles, set per-user spend limits, and review access.

​
Connecting the CLI

The dashboard issues short-lived Bearer tokens that the CLI uses to authenticate. Connect once with: 
synsc connect login
The CLI prints a device code and opens your browser. Confirm the login, and the dashboard issues the token. The CLI stores it in memory (and only in memory) for the session. To verify the connection: 
synsc connect status
Output includes your account email, the credentials currently synced, and your remaining credit balance. To disconnect: 
synsc connect logout
This invalidates the local token. The dashboard credentials stay where they are; the CLI just stops being able to read them until you log in again.

​
Credentials panel

Each provider has its own slot. Paste your token, the dashboard validates the format, and the field is then masked in the UI.
ProviderField
Hugging FaceHF_TOKEN
Weights & BiasesWANDB_API_KEY
ModalMODAL_TOKEN_ID + MODAL_TOKEN_SECRET
Lambda CloudLAMBDA_API_KEY
AWSAWS_ACCESS_KEY_ID + AWS_SECRET_ACCESS_KEY (and optional session token)
GCPService account JSON
Updating a credential is instant, the next session start picks up the new value. To force an immediate refresh in a running session, run synsc connect status again. See Credentials for the in-CLI behavior.

​
Sessions panel

Every time you run synsc, a session record opens in the dashboard. While the session is active, the dashboard shows live model usage and cost. After the session ends, the record persists with:
  • Total tokens (input/output) per model
  • Credits spent
  • Start/end timestamps
  • Working directory the session was launched from
  • Any custom labels you tagged the session with
Use sessions to:
  • Spot loops. A session burning 100k+ tokens against a single prompt is usually stuck.
  • Audit teammates. Team admins can browse sessions across the team to understand spend distribution.
  • Reconstruct work. A session record references the prompts that ran, the model used, and the outcome, useful for compliance or postmortems.

​
Billing panel

The billing panel covers:
  • Plan, your subscription tier and monthly credit grant.
  • Balance, current credits and burn rate.
  • Top-ups, buy additional credits at any time.
  • Invoices, historical billing receipts.
  • Alerts, set thresholds that send a webhook or email when daily/session spend crosses a line.

​
Team management

Team plans support multiple users sharing a credit pool with per-user limits. Admins can:
  • Invite users by email
  • Assign roles (admin, member, read-only)
  • Set per-user daily and monthly spend caps
  • View aggregate and per-user usage
  • Force-disconnect a user (invalidates their CLI Bearer token)
Per-user credentials remain isolated even on a team plan, your AWS key isn’t visible to your teammates.

​
Project tagging

You can tag sessions with a project label so spend rolls up cleanly: 
synsc --project research-fork
The dashboard groups sessions by project so you can answer “how much have we spent on the research-fork project this month” without exporting raw session data.

​
What’s next

  • Credits and usage. The mechanics of pre-call balance checks and per-call reporting.
  • Security. How the dashboard’s RLS, Bearer tokens, and device flow keep credentials and sessions isolated.