Simple, transparent pricing
Start free, scale as you grow. Cloud or self-hosted.
Hosted plans
Five tiers from free Developer to Enterprise. Bring your own LLM key on every plan.
Developer
Full engine for hobby and adoption. Bring your own LLM key.
- • 2 projects
- • 5,000 eval runs / month
- • 10,000 patterns
- • BYOK: OpenAI, Anthropic, Gemini
- • Local embeddings included
- • Community support
Pro
Solo dev in production — analytics, evolution, webhooks.
- • 10 projects
- • 50,000 eval runs / month
- • 100,000 patterns
- • Eval insights + ROI rollups
- • Prompt evolution + failure clustering
- • Outbound webhooks
- • Overage: +$5 / 5,000 runs (opt-in)
Team
RecommendedProduction team — RBAC, audit, async jobs, hosted MCP.
- • 50 projects
- • 250,000 eval runs / month
- • 1,000,000 patterns
- • RBAC (4 roles) + full audit log
- • Async jobs (3 concurrent)
- • Hosted MCP server (Claude Desktop, Cursor)
- • Self-service backup download
Business
Organisation scale — SSO, cross-agent memory, role routing.
- • 250 projects
- • 1,000,000 eval runs / month
- • 10,000,000 patterns
- • SSO / OIDC (Google, Okta, Azure AD)
- • Per-role model routing + provider failover
- • Cross-agent memory pools
- • Quality analytics + 7-year audit retention
Enterprise
Regulated, on-prem, or air-gapped — Vault, SOC 2, signed DPA.
- • Unlimited projects, patterns, eval runs
- • HashiCorp Vault Transit (master key never leaves Vault)
- • On-prem / VPC / air-gapped deploy
- • Signed DPA + SOC 2 control mapping
- • Dedicated SLA 99.95 % + Slack channel
- • IP allowlisting, native Ollama, custom KMS
Run on your own infrastructure
Free for non-commercial use. Commercial self-hosting via enterprise agreement.
Developer License
BSL 1.1 self-hosting for non-commercial use.
- • Source code access
- • JSON or PostgreSQL storage
- • Community support
- • No commercial production use
Enterprise Self-hosted
Commercial self-hosting for regulated or customer-managed environments.
- • Commercial license agreement
- • Air-gapped / VPC deployment
- • SSO, audit log, cross-agent memory
- • Hotfix SLA + dedicated Slack
- • DPA support
Bring your own LLM key
Engramia is the memory and evaluation layer. LLM tokens (OpenAI, Anthropic, others) are billed by you, directly to your provider — we never resell or mark up model usage.
Your key, your bill
Configure your OpenAI, Anthropic, or any supported provider key in the dashboard. Tokens never pass through our billing — you pay the model provider directly.
Eval-runs are a fair-use cap
The per-plan limit is a soft, fair-use guardrail. We will reach out before throttling and discuss higher caps. It is not a hard cliff at the published number.
Your plan price covers Engramia only
Hosting, storage, recall, governance, audit log, and dashboards are included. LLM inference, embeddings, and evaluator-model calls are billed separately by your model provider.