Meet Nexo · Donala guide

Say hi to Nexo.

Ask anything about Donala: persistent AI memory, UniMemory and MultiUser spaces, setup, demos, pricing, or connecting your application.

Try:
1 GiB → ≈177M tokensof additional searchable memory for your AI

One-provider calibration: Gemini 2.5 Flash has a 1,048,576-token window; 1 GiB holds about 169× that as searchable history. Actual database use varies by workload.

The setup

Three steps, then it runs itself

Pricing

Pay only for what you store

01
Connect your memory
Link the PostgreSQL database for this memory from your dashboard.
02
Write every completed turn
Your AI saves the conversation as it works, with no extra work for the user.
03
Every call recalls what matters
Requests use your own API key. Donala retrieves the relevant stored context instead of resending every row.
Growth
25 GB
$29.99/mo

One shared memory pool: every record is shared; do not use it when end-user data must stay separate.

  • Full Nexo agent access
  • Multi-session memory isolation
  • Up to 25 GB memory quota
  • GDPR one-call data release

Memory model

Select or type storage footprint

1 GB (Starter)1024 GB (Enterprise)

What memory includes

Eight capabilities. One product surface.

Storage and confirmed deletion ship with every paid memory. Unlock instructions, schedules, search, personalisation, tagging, and lock when you need them.

Built for product teams shipping agents·Same API from 1 GB to 100 TB·Multiple memories per account

Learn

Video guides

Watch how to set up Donala, connect your database, and build a persistent AI agent in minutes.

Coming soon

Getting started with Donala

Set up your memory space, connect your Postgres DB, and make your first API call in under 5 minutes.

Coming soon

Memory spaces explained

Understand UniMemory vs MultiUser mode and when to use each for your production AI application.

Coming soon

Instructions and scheduling

Lock in persistent AI behaviour with immutable instructions and set up time-based reminders.

Integration

The whole setup is one call

Donala connects to the PostgreSQL database for your memory. Create a memory, connect its database, then add its key to your server-side integration.

  • One API key, one connection string
  • Write and recall in any language
  • Memory persists across restarts
  • GDPR: delete a tenant in one call
curl -X POST https://api.donala.ai/api/memory/write/ \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"id":"msg-001","session_id":"sess-abc",
       "role":"user","content":"Hello","scope":"global"}'

curl -X POST https://api.donala.ai/api/memory/recall/ \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query":"hello","limit":8}'

→ Create a new production key in Dashboard → API Keys, copy it once, then place it in your application's server-only DONALA_API_KEY environment variable. Existing key prefixes are not usable credentials.

Ready?

Your AI finally remembers everything

Start with 1 GB free. Connect your database in under 5 minutes.

balingenensiidan@gmail.com

Contact us

Get in touch directly

Have a specific question, partnership idea, or need a custom Enterprise+ quote?

Talk to Nexo

Nexo
answers in real time
planet · donala
Hi, I'm Nexo. Ask me about pricing, setup, or how memory persistence works here.09:20 AM