Metadata-Version: 2.4
Name: velixar
Version: 1.0.0
Summary: Persistent memory infrastructure for AI applications
Project-URL: Homepage, https://velixarai.com
Project-URL: Documentation, https://docs.velixarai.com
Project-URL: Repository, https://github.com/VelixarAi/velixar-python
Author-email: Velixar <sdk@velixarai.com>
License-Expression: MIT
Keywords: ai,langchain,llm,memory,rag,vector
Classifier: Development Status :: 4 - Beta
Classifier: Intended Audience :: Developers
Classifier: License :: OSI Approved :: MIT License
Classifier: Programming Language :: Python :: 3
Classifier: Programming Language :: Python :: 3.9
Classifier: Programming Language :: Python :: 3.10
Classifier: Programming Language :: Python :: 3.11
Classifier: Programming Language :: Python :: 3.12
Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
Requires-Python: >=3.9
Requires-Dist: httpx>=0.25.0
Requires-Dist: pydantic>=2.0.0
Provides-Extra: all
Requires-Dist: langchain-core>=0.1.0; extra == 'all'
Requires-Dist: llama-index-core>=0.10.0; extra == 'all'
Provides-Extra: dev
Requires-Dist: mypy>=1.0.0; extra == 'dev'
Requires-Dist: pytest-asyncio>=0.21.0; extra == 'dev'
Requires-Dist: pytest-httpx>=0.22.0; extra == 'dev'
Requires-Dist: pytest>=7.0.0; extra == 'dev'
Requires-Dist: ruff>=0.1.0; extra == 'dev'
Provides-Extra: langchain
Requires-Dist: langchain-core>=0.1.0; extra == 'langchain'
Provides-Extra: llamaindex
Requires-Dist: llama-index-core>=0.10.0; extra == 'llamaindex'
Description-Content-Type: text/markdown

# Velixar Python SDK

[![PyPI](https://img.shields.io/pypi/v/velixar)](https://pypi.org/project/velixar/)
[![Python](https://img.shields.io/pypi/pyversions/velixar)](https://pypi.org/project/velixar/)
[![License](https://img.shields.io/github/license/VelixarAi/velixar-python)](LICENSE)

Persistent memory for AI assistants and agents. Give any LLM-powered application long-term recall across sessions.

Velixar is an open memory layer — it works with any AI assistant, agent framework, or LLM pipeline. Store facts, preferences, and context that persist beyond a single conversation.

## Installation

```bash
pip install velixar

# With LangChain integration
pip install velixar[langchain]

# With LlamaIndex integration
pip install velixar[llamaindex]

# All integrations
pip install velixar[all]
```

## Quick Start

```python
from velixar import Velixar

v = Velixar(api_key="vlx_your_key")  # Or set VELIXAR_API_KEY env var

# Store a memory
memory_id = v.store(
    content="User prefers dark mode and metric units",
    tier=0,  # 0=pinned, 1=session, 2=semantic (default), 3=org
    user_id="user_123",
    tags=["preferences"],
)

# Search memories semantically
results = v.search("user preferences", limit=5)
for memory in results.memories:
    print(f"[{memory.score:.2f}] {memory.content}")

# Get context for LLM prompts
context = v.get_context("What are the user's preferences?", max_tokens=2000)
```

## Async Support

```python
from velixar import AsyncVelixar

async with AsyncVelixar(api_key="vlx_...") as v:
    await v.store("User's favorite color is blue", user_id="user_123")
    results = await v.search("favorite color")
```

## Memory Tiers

| Tier | Name | Use Case |
|------|------|----------|
| 0 | Pinned | Critical facts, user preferences, never expire |
| 1 | Session | Current conversation context |
| 2 | Semantic | Long-term memories (default) |
| 3 | Organization | Shared team knowledge (Hivemind+) |

```python
from velixar import MemoryTier

v.store("User is allergic to peanuts", tier=MemoryTier.PINNED)
v.store("Currently discussing project X", tier=MemoryTier.SESSION)
```

## Cognitive Features by Plan

| Feature | Free | Cortex ($29) | Synapse ($75) | Hivemind ($25/seat) |
|---------|------|-------------|----------------|---------------------|
| Store & search | ✓ | ✓ | ✓ | ✓ |
| Neural ensembles | — | ✓ | ✓ | ✓ |
| Temporal chains | — | ✓ | ✓ | ✓ |
| Consolidation | — | ✓ | ✓ | ✓ |
| Identity modeling | — | — | ✓ | ✓ |
| Org memory (tier 3) | — | — | — | ✓ |

Free tier stores and searches memories. Paid tiers activate cognitive features automatically — no code changes needed. [Pricing →](https://velixarai.com/pricing)

## Use With Any AI Assistant

Velixar is assistant-agnostic. Plug it into OpenAI, Anthropic, LangChain, LlamaIndex, custom agents, or any LLM pipeline:

```python
# Inject memories as context before calling your LLM
results = v.search(user_message, limit=5)
context = "\n".join(m.content for m in results.memories)

response = openai.chat.completions.create(
    model="gpt-4",
    messages=[
        {"role": "system", "content": f"Relevant memories:\n{context}"},
        {"role": "user", "content": user_message},
    ],
)

# Store important facts after the conversation
v.store("User prefers concise answers", user_id="user_123")
```

## LangChain Integration

```python
from langchain_openai import ChatOpenAI
from langchain_core.prompts import ChatPromptTemplate, MessagesPlaceholder
from langchain_core.runnables.history import RunnableWithMessageHistory
from velixar.integrations.langchain import VelixarChatMessageHistory

def get_session_history(session_id: str):
    return VelixarChatMessageHistory(session_id=session_id, api_key="vlx_...")

chain = ChatPromptTemplate.from_messages([
    ("system", "You are a helpful assistant."),
    MessagesPlaceholder(variable_name="history"),
    ("human", "{input}"),
]) | ChatOpenAI()

with_history = RunnableWithMessageHistory(
    chain, get_session_history,
    input_messages_key="input",
    history_messages_key="history",
)

# Memory persists across sessions, restarts, and deployments
config = {"configurable": {"session_id": "user_123"}}
with_history.invoke({"input": "I prefer Python over JavaScript"}, config=config)
with_history.invoke({"input": "What language do I prefer?"}, config=config)
```

## LlamaIndex Integration

```python
from llama_index.core.agent import ReActAgent
from llama_index.llms.openai import OpenAI
from velixar.integrations.llamaindex import VelixarMemory

memory = VelixarMemory(api_key="vlx_...", user_id="user_123")
agent = ReActAgent.from_tools(tools=[...], llm=OpenAI(), memory=memory)
```

## Batch Operations

```python
result = v.store_many([
    {"content": "Fact 1", "tier": 0},
    {"content": "Fact 2", "tier": 2, "tags": ["important"]},
    {"content": "Fact 3", "user_id": "user_456"},
])
```

## Error Handling

```python
from velixar import VelixarError, RateLimitError, AuthenticationError

try:
    v.store("test")
except AuthenticationError:
    print("Invalid API key")
except RateLimitError as e:
    print(f"Rate limited. Retry after {e.retry_after}s")
except VelixarError as e:
    print(f"Error: {e.message}")
```

## Configuration

```python
v = Velixar(
    api_key="vlx_...",           # Or VELIXAR_API_KEY env var
    base_url="https://...",      # Custom endpoint (optional)
    timeout=30.0,                # Request timeout in seconds
    max_retries=3,               # Retry attempts for failures
)
```

## Get an API Key

Sign up at [velixarai.com](https://velixarai.com) and generate a key under Settings → API Keys.

## Related

- [velixar (JavaScript SDK)](https://github.com/VelixarAi/velixar-js) — TypeScript/JavaScript client
- [velixar-mcp-server](https://github.com/VelixarAi/velixar-mcp-server) — MCP server for any MCP-compatible AI assistant

## License

MIT
