Hebbrix + Dify
Dify makes it easy to build AI apps with a visual workflow editor. Hebbrix adds the one thing Dify doesn't have: memory that persists across conversations and learns about each user over time.
Dify's conversation memory resets between sessions
Dify uses a token buffer that keeps recent messages in context. It works well within a single conversation. But when a user comes back tomorrow, that context is gone. The app starts from scratch every time.
Dify's knowledge base is great for static documents like FAQs and product specs. But it doesn't learn from conversations. It doesn't remember what a specific user told you last week. That's where Hebbrix comes in.
Three ways to connect
Choose the approach that fits how you're building your Dify app.
External Knowledge API
Dify natively supports connecting external knowledge sources via API. Point it at Hebbrix, and your chatflows and workflows can search memories alongside your document knowledge base. Go to Knowledge, then External Knowledge API in Dify's settings, add your Hebbrix endpoint, and it starts working immediately.
Best for: retrieval focused apps where you want memories in the knowledge pipeline
Custom Tool Plugin
Build a Dify tool plugin that wraps Hebbrix's API. This gives your agents and workflows explicit search and store actions they can call during execution. The agent decides when to store a new memory and when to look one up, giving you full control over the memory lifecycle.
Best for: agent apps that need read and write memory during conversations
External Data Tool
Dify's External Data Tool runs before every LLM call. Connect it to Hebbrix, and relevant memories are automatically injected into every prompt. The user never sees the memory lookup happening. It just works in the background, making every response more personalized.
Best for: transparent personalization without changing your workflow logic
What memory unlocks for Dify apps
Returning User Recognition
Your app knows who's talking. Greet returning users with context about their last conversation, their preferences, and their open questions.
Preference Learning
A user mentions they prefer technical explanations over analogies? Stored. Next time, the app adapts without being asked.
Conversation Continuity
"We were discussing the migration plan yesterday." Your Dify app picks up exactly where you left off because Hebbrix remembers the context.
Connected Context
Hebbrix's knowledge graph connects users to their projects, preferences, and interactions. Ask about one thing and get related context for free.
Temporal Awareness
Memories decay naturally over time. Recent interactions carry more weight than ones from months ago, keeping responses current.
Self-Improving Responses
Which memories led to good responses? Hebbrix tracks this automatically and reinforces the patterns that work.
Dify conversation memory vs. Hebbrix
Dify remembers within a conversation. Hebbrix remembers across all of them.
Give your Dify apps a memory
Free tier. No credit card. Connect to Dify in minutes.