Synapse
Free or Licensed
BYOK only — your provider keys
- Apache-2.0 for community use
- Commercial licence (Synapse EE) — contact for terms
- Bring your own Postgres, SSO, LLM keys
- Memory bounded by your own disk
- Community support (GitHub issues)
Pricing
Memory is the value unit.
Your investment compounds in the one asset you can't easily rebuild: your team's institutional memory. Agents, councils, MCP, integrations, and sandbox-hours all come bundled. Unlimited users at every tier.
Subscription prices below cover platform & memory. LLM token usage is billed separately, and BYOK is available on every plan — bring your own OpenAI / Anthropic / Bedrock / Vertex keys at zero markup, or use Cerebro managed tokens. See token rates →
Cerebro Team
$99 / month
+ LLM tokens — BYOK or managed
- 50 GB memory · unlimited users
- Hosted, managed infrastructure
- BYOK on any provider — zero markup
- OIDC SSO · basic audit log
- All council templates included
- Email support
- $990 / year saves $198
Cerebro Business
$999 / month
+ LLM tokens — BYOK or managed
- 500 GB memory · unlimited users
- Everything in Team, plus:
- SAML SSO · full audit log
- Multi-tenancy
- Mobile push (FCM + APNs)
- BYOK with per-workspace key routing
- Priority support
- $9,990 / year saves $1,998
Cerebro Enterprise
Custom
+ LLM tokens — BYOK or managed
- 5 TB+ memory · multi-org
- Everything in Business, plus:
- SCIM provisioning · DSAR automation
- SOC 2 + GDPR + PDPA + HIPAA
- Dedicated cluster (hosted or on-prem)
- BYOK-only or BYOC supported (regulated)
- Dedicated CSM · 24/7 P1 SLA
- From $100k / year
Why we don't price per seat.
Per-seat pricing taxes your team for growing. Our job is to help you protect and grow your collective intelligence — not to meter the people doing the thinking.
Per-seat tools
Cost grows with hires
- 10 users$150–250 / mo
- 50 users$750–1,250 / mo
- 200 users$3,000–5,000 / mo
- 500 users$7,500–12,500 / mo
Slack Business+ & Microsoft Teams ranges, premium add-ons included. Every hire is a finance conversation.
Cerebro Business
Flat $999 / mo
- 10 users$999 / mo
- 50 users$999 / mo
- 200 users$999 / mo
- 500 users$999 / mo
500 GB memory. Unlimited users. Your team can grow without your bill moving — until your institutional memory outgrows the tier.
Every voice adds judgement — the seasoned advisor, the new hire on day one. None of it should cost extra. Our role is to help your business retain its collective intelligence — so the thinking you build together compounds, no matter who's in the room.
Two ways to pay for LLM tokens.
Pick per workspace. Switch any time.
BYOK — Bring Your Own Key
Connect your OpenAI, Anthropic, Bedrock, or Vertex keys via the Cerebro admin UI. Tokens flow directly to your provider on your account — we never sit in the billing path. Full control over model selection, with spend visible on your provider's existing dashboards.
We add zero markup. Available at every tier including Free. Required for self-hosted Synapse.
Cerebro Tokens — Managed
One relationship, one invoice. Set monthly budgets and per-workspace spend caps — we alert before they're hit, so usage stays on plan. We handle provider selection, key rotation, rate limits, and outages.
Rates improve at Business and Enterprise tiers. BYOK available any time — no lock-in.
Managed token rates · $ per million tokens (input / output)
| Tier | Standard class | Premium class |
|---|---|---|
| Trial | $1.50 / $3.00 | $4.00 / $12.00 |
| Team | $1.20 / $2.50 | $3.50 / $11.00 |
| Business | $1.00 / $2.00 | $3.00 / $10.00 |
| Enterprise | $1.00 / $2.00 | $2.50 / $9.00 |
Up to 87% below Claude Sonnet list pricing on output tokens at Business and Enterprise tiers.
Standard class covers everyday reasoning (GPT-4.1 / Claude Sonnet / Gemini Pro class). Premium covers frontier reasoning (OpenAI o-series, Claude Opus). Reference list: Claude Sonnet 4.6 direct is $3.00 / $15.00 per M; OpenAI GPT-4.1 direct is $2.00 / $8.00 per M. Enterprise customers may also choose BYOK-only — common for regulated on-prem deployments.
Trial is a 30-day workspace to evaluate Cerebro against real workloads — full Team-class features, unlimited seats, your own data. Subscription is $0. LLM token usage is billed at the Trial rates above and hard-capped at $50 over the 30 days — at the cap, managed-token requests pause until you raise the limit, switch to BYOK (uncapped), or convert to Team. Card required so usage can be charged. At day 30, the workspace converts to Team or downgrades to Free (100 MB, BYOK-only) — your call.
Rates indicative as of May 2026 and may be adjusted as upstream provider pricing evolves; contract pricing governs at Business and Enterprise tiers.
No surprise overage cuts.
When a tenant hits the memory cap, recall and reflect operations continue uninterrupted. New memory still writes, billed at a flat $0.20 per GB-month above the tier cap. Email warnings at 80% and 100% of cap. No service interruption.
Overage is a courtesy, not a revenue strategy. Customers consuming significant overage should be having an upgrade conversation — not paying us metered scraps.
Pricing questions
Are LLM token costs included in the subscription?
No. Subscription pricing covers platform and memory. LLM token usage is billed separately. You can bring your own provider keys (BYOK) at zero markup, or use Cerebro managed tokens at tier-discounted rates.
Is BYOK available on every plan?
Yes. BYOK (Bring Your Own Key) is supported on every Cerebro tier — Trial, Team, Business, and Enterprise — with zero markup. Synapse OSS is BYOK by nature.
How does the Trial work?
Trial is a 30-day workspace with $0 subscription and full Team-class features. Managed LLM token usage is hard-capped at $50 total over the 30 days. At day 30, the workspace converts to a paid Team plan or downgrades to Free — your choice, no auto-charge.
Why is pricing based on memory instead of seats?
Per-seat pricing taxes your team for growing. Cerebro is priced on memory capacity — the institutional intelligence accumulated in the platform — because that's the asset that compounds. Every voice adds judgement; none of it should cost extra.
What happens when we hit the memory cap?
Recall and reflect operations continue uninterrupted. New memory writes are billed at $0.20 per GB-month above the tier cap, with email warnings at 80% and 100% of cap. No service interruption.
How do the managed token rates compare to OpenAI / Anthropic direct?
Standard-class output tokens at Business and Enterprise tiers run up to 87% below Claude Sonnet 4.6 list pricing. Reference: Sonnet 4.6 direct is $3.00 / $15.00 per million tokens; our Business Standard rate is $1.00 / $2.00. Personal/Trial Premium is the one place we're a markup on certain frontier reasoning models — for those workloads, BYOK is usually the better path.
Ready to start?
Pick a tier above, or talk to us about the right shape for your team.