Pay for the layer
that compounds.
Solo for the individual dev piping build logs into their CLI. Team for the squad sharing a codebase brain. Enterprise for VPC + SOC 2 + BYOK. Dream Engine, per-task context, and 9 compounding strategies included at every tier.
Opus 4.7, seed=42, temp=0, official swebench 4.1.0 docker evaluator. 30.0% → 40.7% strict, CI [+4.67, +16.67]. Full methodology →
Choose your plan
Kick the tires. Wire the CLI into a side-project repo.
- 1,000 memories
- 50 Dream Engine runs / month
- 10,000 recalls / month
- 1 namespace
- REST API + 1 webhook
- OSS SDKs — Python, Node (Apache-2.0, repos coming Q3 2026)
playground_*namespace only on demo keys
No credit card required.
or $20/month billed annually
The individual dev. One repo, one agent, compounding daily.
- 50,000 memories
- 500 Dream Engine runs / month
- 100,000 recalls / month
- Unlimited namespaces
- All 4 webhook events + WebSocket streams
- MCP server access
- 9 Dream Engine strategies unlocked
- 30-day memory versioning
- Email support via dev@remlabs.ai
or $79/month billed annually
The squad. Shared codebase memory across every agent on the team.
- 500,000 memories
- 5,000 Dream Engine runs / month
- 1M recalls / month
- Everything in Solo
- Multi-agent RBAC (owner/admin/developer/viewer)
- Shared namespaces across team members
- Pub/sub channels
- A2A protocol
- 90-day memory versioning
- Priority support (< 24h response)
Talk to us
VPC + SOC 2 + BYOK for regulated orgs and monorepos.
- Everything in Team
- Unlimited memories / dreams / recalls
- SSO (SAML, OIDC)
- BYOK (AWS KMS / GCP KMS / Azure Key Vault)
- Dedicated VPC deployment or self-hosted (annual license)
- SOC 2 Type I in flight (target Q4 2026) Roadmap
- HIPAA BAA available on Enterprise
- 99.9% uptime SLA with credits (higher per-contract targets negotiable)
- Dedicated Slack / MS Teams channel
- Data residency: US-East live; EU and APAC roadmap (Enterprise dedicated on request)
Overage pricing
Transparent. No surprises.
Hard-cap toggle available in /console to prevent overage (default on for Free + Solo).
What's included in every plan
No feature gates on the fundamentals. Every tier gets the full platform.
Unlimited Storage
We never charge for storing memories
8 Search Modes
Semantic, keyword, hybrid, temporal, and more
Knowledge Graph
Automatic relationship mapping between memories
API Access
Full REST API with SDKs for Python and Node.js
MCP Server
Connect to Claude, Cursor, and any MCP client
Dream Engine is the difference
Most memory systems store and retrieve. The Dream Engine runs nightly consolidation cycles that deduplicate, contradict-detect, and synthesize new structure across your namespace — so the next call returns a denser, cleaner answer than the last.
Free
1 cycle / day
Basic consolidation to keep your memories organized
Solo
Unlimited cycles
All 9 strategies + tournament refinement + scheduled digests
Full feature comparison
| Feature | Free | Solo | Team | Enterprise |
|---|---|---|---|---|
| Memories stored | 10K | 100K | 1M | Unlimited |
| API calls / month | 10K | 100K | 1M | Unlimited |
| Dream Engine cycles | 1 / day | 1 / day | Every 6h | Continuous |
| All 9 strategies | Yes | Yes | Yes | Yes |
| Contradiction detection | — | Yes | Yes | Yes |
| Brain Glow score | Limited | Full | Full (team) | Full (org) |
| Scheduled digest webhook | Weekly | Daily | Daily | Daily |
| Federation | — | — | Yes | Yes |
| RBAC | — | — | Yes | Yes |
| Webhooks | — | Yes | Yes | Yes |
| WebSocket streams | — | Yes | Yes | Yes |
| MCP protocol | Yes | Yes | Yes | Yes |
| Self-host | — | — | — | Yes |
| SSO | — | — | — | Yes |
| Audit logs | — | — | Yes | Yes |
| SLA | — | 99.9% | 99.9% | 99.9% (per-contract higher available) |
| Support | Community | Priority email | Dedicated Slack + phone | |
| Annual invoice | — | — | Yes | Yes |
| SOC 2 | Type I roadmap (Q4 2026) | Type I roadmap (Q4 2026) | Type I roadmap (Q4 2026) | Type I in flight (Q4 2026); Type II planned (Q2 2027) |
| Data residency options | — | — | — | Yes |
| Uptime credits | — | — | Yes | Yes |
| Knowledge graph | Yes | Yes | Yes | Yes |
| Namespaces | 1 | 1 | 10 | Unlimited |
| Tournament refinement | — | Yes | Yes | Yes |
| 8 search modes | Basic | Yes | Yes | Yes |
| Slack/Notion delivery | — | — | Yes | Yes |
| Custom integrations | — | — | — | Yes |
| BAA / DPA | — | — | — | Yes |
| API rate limit | 10 req/s | 50 req/s | 200 req/s | Custom |
| 30-day refund | — | Yes | Yes | Custom |
What plan do I need?
Tell us about your app and we'll recommend a plan.
Pay-as-you-go
For high-volume API users who'd rather pay per operation than per tier.
| Monthly volume | Stores | Recalls | Dreams |
|---|---|---|---|
| 0 – 1M | $0.001 | $0.0005 | $0.01 |
| 1M – 10M | $0.0008 (20% off) | $0.00035 (30% off) | $0.008 |
| 10M+ | Custom (contact) | Custom | Custom |
Automatic tier promotion — your bill applies the best rate retroactively.
Model your monthly bill in 30 seconds.
Three sliders. Real numbers. Based on PAYG rates. No sales call required.
Free forever for universities and OSS maintainers.
Research labs at accredited universities, authors of OSS projects with 1,000+ stars, and published AI research teams qualify for a free Academic tier: all Solo features, unlimited cycles, mention in project acknowledgments. Email dev@remlabs.ai from a .edu or your project's maintainer email with a link.
What will REM save you?
Hours spent re-grounding agents, re-prompting, re-finding context — priced honestly.
3 questions, 1 recommendation
/settings/billing. Your memories stay yours. Download everything as JSON/CSV before closing.What we guarantee, per tier.
The same numbers that go on an Enterprise contract — visible before you talk to sales.
| Commitment | Free | Solo | Team | Enterprise |
|---|---|---|---|---|
| Uptime | Best effort | 99.9% | 99.9% | 99.9% baseline; higher per-contract |
| Support response | Community | 24h email | 4h email | 1h (dedicated Slack) |
| Data backup | Daily | Daily | Hourly | Hourly + geo-replicated |
| Incident credit | — | 10% for <99.9% | 10% for <99.9% | Up to 100% per contract |
| Security review | Public docs | Public docs | Questionnaire | SOC 2 Type I in flight (Q4 2026), pentest Q3 2026, DPA |
| Data residency | us-east-1 | us-east-1 | us-east-1 | us-east-1 live; EU/US-West/APAC roadmap, dedicated on request |
See /sla for the full terms, credit math, and incident-reporting procedure.
Common questions
Yes — upgrade instantly, prorated to the day. Your new limits apply immediately; you pay only the difference for the remaining days in your current billing cycle.
The Free tier is our trial. It includes all 9 Dream Engine strategies and MCP protocol so you can evaluate the core product end-to-end. When you're ready for higher limits, Solo is $25/mo ($20 annual) with a 30-day money-back guarantee.
One stored record: a fact, preference, message, or chunk you wrote via memory.set() or batch(). Vector + graph edges are included free. Compressed/consolidated outputs from Dream cycles do not count against your limit.
No. 4xx and 5xx responses don't count against your API call quota. Rate limits (429) also don't count. You only pay for operations that actually executed on your data.
Yes, on Team and Enterprise. Plug in Claude, GPT-4o, Llama/Mixtral via Together/Fireworks/Groq, or a self-hosted Ollama endpoint. Dream Engine runs the 9 strategies against the model you configure, and we pass through token costs at provider price.
USDC (Base, Ethereum, Solana) is supported on annual Team and Enterprise plans — invoice-based, not self-serve. Email billing@remlabs.ai to set up.
30-day money-back on monthly plans, no questions asked. Annual plans are pro-rated to the unused months. Enterprise contracts follow the terms in your MSA.
Yes. 501(c)(3) non-profits and accredited academic labs get 50% off Solo and Team indefinitely; students with a .edu address get Solo free for 12 months. Email dev@remlabs.ai with verification.
We notify you at 80% usage. There is no hard cutoff — your app never goes down. You can upgrade instantly or we soft-limit until your next billing cycle.