The memory infrastructure
for AI agents
One API for four memory types. Working, semantic, episodic, and procedural — all serverless, all AWS-native. Give your agents persistent memory in minutes, not months.
Built for your stack
Your first agent memory in 15 lines
Install pip install mnemora and you're ready.
1from mnemora import MnemoraSync23with MnemoraSync(api_key="mnm_...") as client:4 # Store working-memory state5 client.store_state("agent-1", {"task": "summarize Q4", "step": 1})67 # Semantic memory — auto-embedded server-side8 client.store_memory("agent-1", "User prefers bullet points over prose.")910 # Vector search across all stored memories11 results = client.search_memory("user formatting preferences", agent_id="agent-1")12 for r in results:13 print(r.content, r.similarity_score)1415 # Log an episode to the time-series history16 client.store_episode(agent_id="agent-1", session_id="sess-001",17 type="action", content={"tool": "summarize", "input": "Q4 report"})
AI agents are stateless by default
To give them memory, developers stitch together four databases, four clients, four billing accounts — with no unified query layer.
Four databases to stitch together
Redis for state. Pinecone for vectors. Postgres for structured data. S3 for logs. Four billing accounts, four clients, zero unified memory layer.
No unified query layer
State lives here. Vectors live there. Events are somewhere else. You write glue code instead of writing agents.
Agents forget everything
Without persistence, every invocation starts from scratch. Context, preferences, and history vanish between calls.
One API. Four memory types.
Serverless-first. Every component scales to zero. No LLM required for CRUD operations.
Working Memory
Key-value state in DynamoDB. Sub-10ms reads with optimistic locking and configurable TTL. Ideal for agent step state and task context.
Semantic Memory
Natural-language text stored as 1024-dimensional vectors in Aurora pgvector. Auto-embedded via Bedrock Titan. Duplicates are merged, not re-inserted.
Episodic Memory
Append-only time-series event log. Hot data in DynamoDB, automatically tiered to S3. Full session replay and time-range queries.
Procedural Memory
Tool definitions, prompt templates, schemas, and rules stored in Postgres. Version-controlled and queryable by name. Schema is live; SDK methods ship in v0.2.
How Mnemora compares
Concrete data. No hype.
| Feature | Mnemora | Mem0 | Zep | Letta |
|---|---|---|---|---|
| Memory types | 4 (state, semantic, episodic, procedural) | 1 (semantic only) | 2 (semantic + temporal) | 2 (core + archival) |
| Vector search | pgvector 1024d | External DB | Built-in | Built-in |
| LLM required for CRUD | Every op | Every op | ||
| Serverless | ||||
| Self-hostable | Partial | |||
| Multi-tenant | ||||
| LangGraph checkpoints | ||||
| State latency | <10ms | ~500ms | <200ms | ~1s |
Data based on public documentation as of 2025. Subject to change.
From generic responses to personalized support
overall quality improvement
Measured across 10 real customer scenarios with HubSpot CRM data
Customer asks: “I spoke to someone last week about my API issue, any updates?”
“I'd be happy to help with your API integration issue. Could you provide your account email or ticket number? What specific issue were you experiencing?”
“I can see you've had a few API-related items with us recently. Are you referring to the API rate limiting issue where you were getting 429 errors in production? That's currently marked as high priority and still being worked on.”
Powered by HubSpot CRM data synced to Mnemora semantic memory
Built for every agent type
From customer support to autonomous research — Mnemora adapts to your agent's memory needs.
Customer Support Agent
Remembers every customer interaction, surfaces relevant past tickets, and personalizes responses across channels.
Episodic memory logs each conversation. Semantic search finds similar past issues to suggest resolutions.
Research Agent
Accumulates findings across sessions, deduplicates sources, and builds a searchable knowledge base over time.
Semantic memory stores and deduplicates research findings. Episodic logs track the full research trail.
Coding Assistant
Maintains project context, remembers architectural decisions, and tracks code review patterns per repository.
Working state holds active context. Semantic memory recalls past decisions and patterns across sessions.
Sales Agent / CRM
Tracks deal stages, remembers client preferences, and surfaces talking points from previous meetings.
State tracks deal pipeline status. Episodic memory replays past interactions for personalized follow-ups.
Workflow Automation
Persists multi-step workflow state with optimistic locking, checkpoints progress, and resumes on failure.
LangGraph checkpoints persist workflow state. Procedural memory stores rules and tool definitions.
AI Tutor
Adapts to each learner, tracks mastery across topics, and builds a personalized curriculum over time.
Episodic memory tracks learning history. Semantic memory stores curriculum knowledge for adaptive recall.
Simple, transparent pricing
Start free. Scale as you grow. No surprises.
Free
For exploration and side projects
- 500 API calls / day
- 1 agent
- 50 MB storage
- 5K vectors
- Community support
Starter
For early-stage products
- 5,000 API calls / day
- 10 agents
- 500 MB storage
- 50K vectors
- Email support
Pro
For production applications
- 25,000 API calls / day
- 50 agents
- 5 GB storage
- 250K vectors
- Priority support
Scale
For high-volume teams
- 50,000 API calls / day
- Unlimited agents
- 10 GB storage
- 500K vectors
- Dedicated support
Enterprise
Unlimited everything, custom SLA, SSO/SAML, VPC peering, and dedicated infrastructure for organizations with custom needs.
All plans include TLS encryption, AWS-native infrastructure, all four memory types, and the full Python SDK. No credit card required for Free.
Frequently asked questions
Ready to give your agents memory?
Start in under 5 minutes. No infrastructure to configure. No servers to manage. Just memory that works.