Sovant
Dashboard →

Quickstart Guide

Quickstart Guide

Get started with Sovant in 5 minutes.

1. Get Your API Key

Sign up at sovant.ai and get your API key from the dashboard.

Your key will look like: sk_live_22955ddcb441e098d23599d32cb390e5...

2. Install SDK

JavaScript/TypeScript

npm install @sovant/sdk

Python

pip install sovant

cURL

No installation needed - use your terminal directly.

3. Initialize Client

JavaScript

import { Sovant } from "@sovant/sdk";

const sv = new Sovant({
  apiKey: process.env.SOVANT_API_KEY,
});

Python

import os
from sovant import Sovant

client = Sovant(api_key=os.environ["SOVANT_API_KEY"])

4. Create Your First Memory

JavaScript

const memory = await sv.memory.create({
  data: "User prefers flat white coffee",
  type: "preference",
  tags: ["beverage", "coffee"],
});

console.log("Created memory:", memory.id);

Python

from sovant import MemoryCreate

memory = client.memory_create(MemoryCreate(
    data="User prefers flat white coffee",
    type="preference",
    tags=["beverage", "coffee"],
))

print(f"Created memory: {memory['id']}")

cURL

curl -X POST https://sovant.ai/api/v1/memory \
  -H "Authorization: Bearer sk_live_your_api_key" \
  -H "Content-Type: application/json" \
  -d '{
    "content": "User prefers flat white coffee",
    "type": "preference",
    "tags": ["beverage", "coffee"]
  }'

Note: The SDK accepts data which it converts to content for the API. When using cURL/REST directly, send content.

5. Search Memories

JavaScript

const results = await sv.memory.search({
  query: "coffee preferences",
  limit: 5,
});

for (const r of results.results) {
  console.log(r.content);
}

Python

from sovant import SearchQuery

results = client.memory_search(SearchQuery(
    query="coffee preferences",
    limit=5,
))

for r in results["results"]:
    print(r["content"])

6. Recall (Hybrid Search)

Recall uses Sovant's hybrid pipeline (profile fast-path + lexical + semantic) for conversational queries like "What do you know about me?".

JavaScript

const recall = await sv.memory.recall({
  query: "coffee preferences",
  limit: 5,
});

for (const r of recall.results) {
  console.log(r.content);
}

Python

recall = client.memory_recall(
    query="coffee preferences",
    limit=5,
)

for r in recall.get("results", []):
    print(r["content"])

7. Threads (Optional)

Threads let you organize related memories into conversations or sessions. Both SDKs have full thread support.

JavaScript

const thread = await sv.threads.create({
  title: "Coffee Preferences",
});

await sv.memory.create({
  data: "User likes oat milk in their coffee",
  type: "preference",
  thread_id: thread.id,
});

const threadRecall = await sv.memory.recall({
  query: "milk preference",
  thread_id: thread.id,
});
console.log(threadRecall.results);

Python

thread = client.threads_create(
    title="Coffee Preferences",
)

client.memory_create(MemoryCreate(
    data="User likes oat milk in their coffee",
    type="preference",
    thread_id=thread["id"],
))

thread_recall = client.memory_recall(
    query="milk preference",
    thread_id=thread["id"],
)
print(thread_recall.get("results", []))

Next Steps