Neura Python SDK

External brain for AI agents — persistent memory and state via HTTP.

from neura import Neura

neura = Neura(api_key="sk-...")

# Store a memory
neura.memory.create(content="User prefers dark mode", tags=["preference"])

# Semantic search
results = neura.memory.search("What are my preferences?")

# Persistent state
neura.state.set("current_goal", {"task": "Build the API"})
goal = neura.state.get("current_goal")

Install

pip install neura-api

Usage

Memory

# Create
mem = neura.memory.create(
    content="The user likes brutalist design",
    tags=["preference", "ui"],
    importance=8,
    metadata={"source": "conversation"},
)

# Semantic search
results = neura.memory.search("design preferences")

# Advanced search with filters
filtered = neura.memory.search_advanced(
    query="user settings",
    filters={"tags": ["preference"], "importance_min": 5},
    limit=20,
)

# Update
neura.memory.update(mem["id"], importance=9)

# Delete
neura.memory.delete(mem["id"])

State

# Set
neura.state.set("risk_level", "conservative")

# Get
risk = neura.state.get("risk_level")
# {"key": "risk_level", "value": "conservative", ...}

# List all
all_state = neura.state.list()

# Delete
neura.state.delete("risk_level")

Error Handling

from neura import NeuraHttpError

try:
    neura.memory.search("something")
except NeuraHttpError as e:
    print(e.code)         # 'rate_limited', 'not_found', etc.
    print(e.message)      # Human-readable
    print(e.action)       # 'wait_and_retry', 'check_api_key', etc.
    print(e.retry_after)  # Seconds to wait

Retries

The SDK automatically retries on network errors and rate limits with exponential backoff.

API