Deja Vu is currently in beta. API access is live. We're now accepting Venice (VVV) as payment, with a discount for early users during the beta period.

Close

Deja Vu is currently in beta. API access is live. We're now accepting Venice (VVV) as payment, with a discount for early users during the beta period.

Close

About

Features

How it works

FAQ

Team

Github

About

Features

How it works

FAQ

Team

Github

Github

Github

Github

Your AI Memory, Owned by You

Your AI Memory, Owned by You

Your AI Memory, Owned by You

Private memory across every AI you use. Powered by Venice. Stored on your device. No cloud. No harvesting. Just context that's always with you.

Github

Github

10M+

10M+

10M+

10M+

10M+

10M+

MEMORY OPERATIONS PROCESSED

44,000+

44,000+

44,000+

44,000+

44,000+

44,000+

GITHUB STARS ON THE ENGINE INSIDE

0 BYTES

0 BYTES

0 BYTES

0 BYTES

0 BYTES

0 BYTES

OF YOUR DATA RETAINED

WHY DEJA VU?

Works across every AI

01/05

Powered by Venice

Every memory Deja Vu extracts is processed through Venice, a decentralized inference network with zero data retention. Your conversations are never logged, stored, or used for training. This is enforced by architecture, not policy.

Learn more

Works across every AI

02/05

Works Across Every AI

ChatGPT, Claude, Gemini, Cursor, Perplexity, open-source models. Deja Vu connects to all of them through MCP, a browser extension, and a local API. Your context travels with you no matter which tool you use.

Stored on your device

03/05

Stored on Your Device

Your memories live in a local database file on your machine. Not in the cloud. Not behind an API. You can open it, read it, edit it, delete anything, and back it up whenever you want. If Deja Vu disappeared tomorrow, your data stays.

Gets smarter over time

04/05

Gets Smarter Over Time

Day one, it knows your name. After a month, it has a detailed map of your projects, preferences, and communication style. Every AI conversation improves because every AI already has your context. That compounding intelligence is yours.

Production-grade memory engine

05/05

Production-Grade Memory Engine

Deja Vu's extraction pipeline is built on the most widely adopted open-source memory architecture in AI. Entity recognition, fact deduplication, conflict resolution, and multi-signal retrieval, refined across hundreds of millions of real-world operations.

Powered by Venice

01/05

Powered by Venice

Every memory Deja Vu extracts is processed through Venice, a decentralized inference network with zero data retention. Your conversations are never logged, stored, or used for training. This is enforced by architecture, not policy.

Learn more

Works across every AI

02/05

Works Across Every AI

ChatGPT, Claude, Gemini, Cursor, Perplexity, open-source models. Deja Vu connects to all of them through MCP, a browser extension, and a local API. Your context travels with you no matter which tool you use.

Stored on your device

03/05

Stored on Your Device

Your memories live in a local database file on your machine. Not in the cloud. Not behind an API. You can open it, read it, edit it, delete anything, and back it up whenever you want. If Deja Vu disappeared tomorrow, your data stays.

Gets smarter over time

04/05

Gets Smarter Over Time

Day one, it knows your name. After a month, it has a detailed map of your projects, preferences, and communication style. Every AI conversation improves because every AI already has your context. That compounding intelligence is yours.

Production-grade memory engine

05/05

Production-Grade Memory Engine

Deja Vu's extraction pipeline is built on the most widely adopted open-source memory architecture in AI. Entity recognition, fact deduplication, conflict resolution, and multi-signal retrieval, refined across hundreds of millions of real-world operations.

Powered by Venice

01/05

Powered by Venice

Every memory Deja Vu extracts is processed through Venice, a decentralized inference network with zero data retention. Your conversations are never logged, stored, or used for training. This is enforced by architecture, not policy.

Learn more

Works across every AI

02/05

Works Across Every AI

ChatGPT, Claude, Gemini, Cursor, Perplexity, open-source models. Deja Vu connects to all of them through MCP, a browser extension, and a local API. Your context travels with you no matter which tool you use.

Stored on your device

03/05

Stored on Your Device

Your memories live in a local database file on your machine. Not in the cloud. Not behind an API. You can open it, read it, edit it, delete anything, and back it up whenever you want. If Deja Vu disappeared tomorrow, your data stays.

Gets smarter over time

04/05

Gets Smarter Over Time

Day one, it knows your name. After a month, it has a detailed map of your projects, preferences, and communication style. Every AI conversation improves because every AI already has your context. That compounding intelligence is yours.

Production-grade memory engine

05/05

Production-Grade Memory Engine

Deja Vu's extraction pipeline is built on the most widely adopted open-source memory architecture in AI. Entity recognition, fact deduplication, conflict resolution, and multi-signal retrieval, refined across hundreds of millions of real-world operations.

FEATURES

01

Memory that follows you everywhere

Install Deja Vu and store memories locally

Deja Vu connects to your AI tools through MCP and a browser extension. Use Claude for code, ChatGPT for writing, Gemini for research. Your context is already in each one, no re-explaining needed.

02

Venice-powered privacy

Encrypted Venice routing with zero retention — privacy shield

The AI that reads your conversations to decide what to remember runs entirely through Venice. Encrypted routing, decentralized GPU nodes, zero retention. The most sensitive step in the pipeline produces no persistent data anywhere.

03

Two-minute setup

Install Deja Vu and run from the terminal in minutes

Install with pip. Paste your Venice API key. Run the server. Your AI tools now have memory. Memories accumulate in a local file at ~/.dejavu/memories.db. Open it anytime. It's yours.

HOW DEJA VU WORKS

How Novyra
Works

01

Install Deja Vu

Run pip install dejavu-memory and paste your Venice API key. Configuration saves locally. No accounts, no cloud signup, no third-party dependencies.

01

Install Deja Vu

Run pip install dejavu-memory and paste your Venice API key. Configuration saves locally. No accounts, no cloud signup, no third-party dependencies.

01

Install Deja Vu

Run pip install dejavu-memory and paste your Venice API key. Configuration saves locally. No accounts, no cloud signup, no third-party dependencies.

02

Use AI like normal

Talk to ChatGPT, Claude, Gemini, or any AI tool. Deja Vu's browser extension and MCP server observe your conversations in the background.

02

Use AI like normal

Talk to ChatGPT, Claude, Gemini, or any AI tool. Deja Vu's browser extension and MCP server observe your conversations in the background.

02

Use AI like normal

Talk to ChatGPT, Claude, Gemini, or any AI tool. Deja Vu's browser extension and MCP server observe your conversations in the background.

03

Memories are extracted privately

Deja Vu sends your conversations through Venice's zero-retention inference to identify facts, preferences, and context worth remembering. Venice processes the request and forgets it immediately.

03

Memories are extracted privately

Deja Vu sends your conversations through Venice's zero-retention inference to identify facts, preferences, and context worth remembering. Venice processes the request and forgets it immediately.

03

Memories are extracted privately

Deja Vu sends your conversations through Venice's zero-retention inference to identify facts, preferences, and context worth remembering. Venice processes the request and forgets it immediately.

04

Context is stored locally

Extracted memories are saved to a SQLite database on your device. You control what stays, what gets edited, and what gets deleted. No one else has access.

04

Context is stored locally

Extracted memories are saved to a SQLite database on your device. You control what stays, what gets edited, and what gets deleted. No one else has access.

04

Context is stored locally

Extracted memories are saved to a SQLite database on your device. You control what stays, what gets edited, and what gets deleted. No one else has access.

05

Your AI already knows you

Next time you open any AI tool, Deja Vu injects relevant context automatically. You never re-explain yourself again.

05

Your AI already knows you

Next time you open any AI tool, Deja Vu injects relevant context automatically. You never re-explain yourself again.

05

Your AI already knows you

Next time you open any AI tool, Deja Vu injects relevant context automatically. You never re-explain yourself again.

WHAT CHANGES WHEN AI REMEMBERS YOU

WHAT CHANGES WHEN AI REMEMBERS YOU

No more starting over

Open any AI tool and it already knows your projects, your role, your preferences. No preamble. No re-explaining. You pick up right where you left off, even if you've never used that tool before.

No more starting over

Open any AI tool and it already knows your projects, your role, your preferences. No preamble. No re-explaining. You pick up right where you left off, even if you've never used that tool before.

No more starting over

Open any AI tool and it already knows your projects, your role, your preferences. No preamble. No re-explaining. You pick up right where you left off, even if you've never used that tool before.

Switch tools freely

Try the latest model from any provider without losing a thing. Your memory is yours, not theirs. There's no cost to switching and no reason to stay locked into one platform.

Switch tools freely

Try the latest model from any provider without losing a thing. Your memory is yours, not theirs. There's no cost to switching and no reason to stay locked into one platform.

Switch tools freely

Try the latest model from any provider without losing a thing. Your memory is yours, not theirs. There's no cost to switching and no reason to stay locked into one platform.

Your private conversations stay private

Ask about anything without wondering who's reading it on the other end. Health, finances, legal questions, business strategy. Venice processes and forgets. Your device holds the rest.

Your private conversations stay private

Ask about anything without wondering who's reading it on the other end. Health, finances, legal questions, business strategy. Venice processes and forgets. Your device holds the rest.

Your private conversations stay private

Ask about anything without wondering who's reading it on the other end. Health, finances, legal questions, business strategy. Venice processes and forgets. Your device holds the rest.

AI that actually gets you

The more you use it, the less you have to say. Deja Vu learns how you think, how you write, and what you care about. Over time, every AI interaction feels like talking to someone who already knows the context.

AI that actually gets you

The more you use it, the less you have to say. Deja Vu learns how you think, how you write, and what you care about. Over time, every AI interaction feels like talking to someone who already knows the context.

AI that actually gets you

The more you use it, the less you have to say. Deja Vu learns how you think, how you write, and what you care about. Over time, every AI interaction feels like talking to someone who already knows the context.

The future of AI is personal.

Your memory. Your device. Powered by Venice.

Your memory. Your device. Powered by Venice.

Github

Github

Get in touch

Get your API key.

Get your API key.

Get your API key.

Private AI memory across every platform. Powered by Venice.

Sign up to receive news, updates, and release notes.

By submitting this form, you consent to the processing of your personal data in accordance with our Privacy Policy.

Terms of Use

Privacy Policy

Cookie Policy

id="overlay">