Cognitive Layer That Makes Your AI Model Smarter Over Time
AuraSDK
Cognitive Layer That Makes Your AI Model Smarter Over TimeSelf-learning · No fine-tuning · No cloud training · <1ms recall · ~3 MB
Your AI model is smart. But it forgets everything after every conversation — and it never gets smarter from experience.
AuraSDK is a cognitive layer that runs alongside any LLM. It observes every interaction, builds beliefs from patterns, discovers causal relationships, and derives behavioral policies — all locally, without fine-tuning or cloud training. The longer it runs, the smarter your agent becomes.
pip install aura-memory
from aura import Aura, Level
brain = Aura("./agent_memory")
brain.enable_full_cognitive_stack() # activate all 5 cognitive layers
# store what happens
brain.store("User always deploys to staging first", level=Level.Domain, tags=["workflow"])
brain.store("Staging deploy prevented 3 production incidents", level=Level.Domain, tags=["workflow"])
# recall — cognitive layer automatically surfaces relevant patterns
context = brain.recall("deployment decision") # <1ms, no API call
# after enough interactions, the system derives this on its own:
hints = brain.get_surfaced_policy_hints()
# → [{"action": "Prefer", "domain": "workflow", "description": "deploy to staging first"}]
No API keys. No embeddings. No cloud. The model stays the same — your agent gets smarter.
⭐ If AuraSDK is useful to you, a GitHub star helps us get funding to continue development from Kyiv.
Why Aura?
| Aura | Mem0 | Zep | Cognee | Letta/MemGPT | |
|---|---|---|---|---|---|
| Architecture | 5-layer cognitive engine | Vector + LLM | Vector + LLM | Graph + LLM | LLM orchestration |
| Self-learning without LLM | Yes — Belief→Causal→Policy | No | No | No | No |
| Behavioral policies from experience | Yes — automatic | No | No | No | No |
| LLM required | No | Yes | Yes | Yes | Yes |
| Recall latency | <1ms | ~200ms+ | ~200ms | LLM-bound | LLM-bound |
| Works offline | Fully | Partial | No | No | With local LLM |
| Cost per operation | $0 | API billing | Credit-based | LLM + DB cost | LLM cost |
| Binary size | ~3 MB | ~50 MB+ | Cloud service | Heavy (Neo4j+) | Python pkg |
| Memory decay & promotion | Built-in | Via LLM | Via LLM | No | Via LLM |
| Trust & provenance | Built-in | No | No | No | No |
| Encryption at rest | ChaCha20 + Argon2 | No | No | No | No |
| Language | Rust | Python | Proprietary | Python | Python |
The Core Idea: Cheap Model + Aura > Expensive Model Alone
Fine-tuning costs thousands of dollars and weeks of work. RAG requires embeddings and a vector database. Context windows are expensive per token.
Aura gives you a third path: a cognitive layer that accumulates experience between conversations — free, local, sub-millisecond.
Week 1: GPT-4o-mini + Aura Week 1: GPT-4 alone
→ average answers → average answers
Week 4: GPT-4o-mini + Aura Week 4: GPT-4 alone
→ knows your workflow → still forgets everything
→ surfaces patterns you repeat → same cost per token
→ warns before risky actions
Tools (8)
recallSurfaces relevant patterns from memory with sub-millisecond latency.recall_structuredRetrieves structured memory data based on specific criteria.storeStores information into the cognitive memory layer.store_codeStores code snippets into the cognitive memory layer.store_decisionStores a decision-making event for future pattern recognition.searchPerforms a search across stored cognitive memories.insightsRetrieves derived insights from accumulated memory patterns.consolidateTriggers the consolidation of memory layers to refine behavioral policies.Configuration
{"mcpServers": {"aura-memory": {"command": "python", "args": ["-m", "aura.mcp"]}}}