Long-term memory for AI agents that organizes itself and corrects itself
Stage
Not given
MVP
Founded
Not given
Not given
Based in
Not given
SE
Pricing model
Freemium
Free
Pricing
Serverless inference with per-token pricing (pay for output, input, and cached tokens). Training pricing from $0.50 to $10.00 per 1M training tokens depending on model size and fine-tuning type. On-demand deployments at $0.134 to $0.334 per minute ($8 to $20 per hour) for various GPU types.
Personal: Free during early access. Unlimited memories, all agents with separate memories, no card required. Enterprise: On request, set up by the team for departmental organization with cross-department sharing and direct support.
Key features
Guided training runs
Configuration-led training
Custom training logic
Serverless inference
On-demand dedicated deployments
Reserved capacity
OpenAI compatible API
Model library with latest open models
Reinforcement learning support
Multi-region deployments
Self-correcting memory that updates when facts change
Historical memory queries with time-based retrieval
Automatic topic organization and hierarchy
Separate memory for each agent with controlled sharing
Multi-agent support via MCP, REST API, or LangGraph
AES-256 encryption at rest and in backups
OAuth 2.1 authentication and API key management
Data export and deletion capabilities
37 MCP tools with automated verification on every code change
What makes it different
Drop-in replacement for closed-model APIs with cost savings of 50-75%
Instant production deployment from any training checkpoint in seconds
Optimized inference engine for industry-leading throughput and latency
Elastic and global RL inference scaling
Own the complete learning loop for specialized intelligence
Understands what changed, when it was true, and where it belongs - beyond vector database similarity
Never contradicts itself by automatically replacing outdated facts
Fully GDPR aligned with built-in privacy by design
Two-minute integration with Claude, ChatGPT, Cursor and any MCP client
Data stays in Europe on AWS eu-north-1 and never used for model training