Fflidget.com
Flidget
“Databaset let us add memory to Flidget's retention copilot without a vector DB side quest. We store user context once and recall it on every exit flow.”
Vishal Chaudhary
Founder, Flidget
AI memory API for production apps
AI that remembers what matters.
Databaset stores user context semantically and delivers the right memory to the LLM, every time.
Start with 3,000 API calls in your first month. No credit card required.
Privacy FirstSecure by DesignBuilt for AI Developers
Semantic storage
Store user context once. Databaset indexes it for meaning, not keywords.
Instant recall
Fetch the right memories before every LLM call. Sub-50ms target latency.
Any model
Works with GPT, Claude, Gemini, LangChain, and Vercel AI SDK.
Production ready
Dashboard, API keys, usage limits, and billing built in from day one.
3K
API calls month 1
<50ms
recall target
2
official SDKs
USER
PostgreSQL + dark mode
toolpreference...
SEMANTIC UNDERSTANDING
DATABASE
MEMORY LAYER
INSTANT RECALL
LATER
What stack should I use?
questionseeking advice...
AI ANSWER
Since you prefer PostgreSQL and dark mode...
Integrations
Works with every AI model and framework.
From side projects to production AI teams.
OpenAI
Claude
Gemini
LangChain
Vercel AI SDK
Next.js
OpenAI
Claude
Gemini
LangChain
Vercel AI SDK
Next.js
OpenAI
Claude
Gemini
LangChain
Vercel AI SDK
Next.js
STORE. RECALL. SHIP.
Built for modern AI product teams
Databaset is built on the habits that make the best AI teams successful: ship memory fast, keep retrieval semantic, and never rebuild vector plumbing per project.
store(userId, text)
await memory.store({
userId: "u_42",
text: "Prefers PostgreSQL"
})
Store. Save conversations, preferences, documents, notes, and events as persistent memory for AI.
recall(userId, query)
const { memories } = await memory.recall({
userId: "u_42",
query: userMessage
})
Recall. Find memories by meaning, not keywords.
inject into prompt
system += memories.map(m => m.text)
await openai.chat.completions.create(...)
Inject. Use recalled context with GPT, Claude, Gemini, or any model.
Quickstart
Four steps to production memory
Install, store, recall, inject. No vector database project required.
1
Install
One package. No other dependencies.
npm install @databaset/sdk
2
Store
Send any text. We handle chunking, embedding, and storage automatically.
await memory.store(userId, "User prefers dark mode")
3
Recall
Semantic search returns the most relevant memories.
const context = await memory.recall(userId, userMessage)
4
Use with any AI
Inject context into Claude, GPT-4, Gemini, or any model.
const reply = await anthropic.messages.create({
system: `You know this about the user: ${context}`,
messages: [{ role: 'user', content: userMessage }]
})THREE LINES. PRODUCTION READY.
Three lines. Production ready.
Developers do not want another infrastructure project.
Install the SDK, store a memory, recall context, and reply. That is the full AI agent memory loop.
await memory.store(...) const context = await memory.recall(...) reply(context)
That is it.
Read the quickstart →typescript
import { Memory } from '@databaset/sdk' const memory = new Memory() await memory.store({ userId: 'user_123', text: 'User prefers dark mode and VS Code', metadata: { source: 'onboarding' },})Install
npm i @databaset/sdk
Latency target
<50ms p95 recall
Models
GPT · Claude · Gemini
Platform
One memory layer. Ship faster.
Everything you need to add long-term memory to AI products without standing up a vector stack.
Semantic recall built in.
Hybrid vector + keyword search. No Pinecone, no chunk tuning, no embedding pipeline to maintain.recall → 3 memories
• PostgreSQL preference
• dark mode UI
• side project stack
Replace the memory stack.
No separate vector DB, embedding service, or retrieval layer for user context.PineconeEmbeddingsChunkingRAG
Dashboard included.
Inspect memories, usage, and API keys from day one on every plan.SDKs ready.
Node.js and Python SDKs with store, recall, and forget in a few lines.npm i @databaset/sdk
await memory.store({...})Production limits.
Per-org rate limits, monthly quotas, and billing that scales with your app.3K calls in month 1
TRY IT LIVE
Live demo
Try the AI memory API live
Save up to 3 memories, ask 3 questions, and watch Databaset recall context before answering. No signup to start.
Demo: 0/3 memories·0/3 questions
StoreRecallAnswer
Memories0/3
Stored
- Loading...
Add up to 3 facts on the left, then ask me up to 3 questions. I'll recall what matters before I answer.
Testimonials
Trusted by founders shipping real products
Founders at Flidget, PRBoard, and MedOn use Databaset for long-term AI memory without standing up their own vector stack.
Pricing
Start free with your whole team. Upgrade when you need more API calls, apps, and support.
Free
Perfect for side projects and prototypes
$0/month
- 3,000 API calls in your first month
- Full API access to integrate and test
- 1 app
- 7 day retention
- Community support
- Upgrade for production after month one
Most Popular
Starter
For production apps with real users
$29/month
- ≈ 100,000 active users worth of memory storage
- 50,000 recalls per month
- 10 apps
- Unlimited retention
- Email support
- Dashboard analytics
- Webhooks
Growth
For teams scaling AI products
$99/month
- ≈ 1M active users worth of memory storage
- 500,000 recalls per month
- Unlimited apps
- Unlimited retention
- Priority support
- Advanced analytics
- Custom metadata
- SLA guarantee
Enterprise
Custom limits and dedicated support
Custom
- Unlimited everything
- Self-hosted deployment option
- Dedicated support engineer
- Custom contract and billing
- SOC2 compliance
- HIPAA available
Got questions?
If you cannot find what you are looking for, get in touch.