Assistela Docker

Your own server or NAS

Run Assistela on your own Linux server or a compatible NAS, with a browser interface and the free iOS Client.

Chat, memory, files, tools and integrations are free. Assistela Pro adds Personal Assistant and Scheduler. Assistela Client is free.

What you need

Linux ARM64 or AMD64 with Docker Engine and Docker Compose.
Persistent writable storage for the workspace, settings and backups.
An AI model endpoint reachable from the container.
An internet connection for account services and device pairing.

Local or hosted AI

Connect LM Studio or an OpenAI-compatible model server, or choose a supported AI provider. The model runs separately from Assistela; any provider charges are separate from Pro.

Use a model address reachable from Docker. Localhost inside a container refers to that container, not to the computer running your model.

Your data, your host

Docker stores the workspace in a persistent /data volume on your host. Back up this volume and keep the host available. Data needed for a request may be sent to the AI provider you choose.

Platform differences

Docker provides server integrations and a Chromium browser. It does not control Mac apps or directly access Apple Mail, iMessage or Apple’s local calendars and contacts.

Installation and first connection

Download the ready-to-run image from GitHub. Docker selects ARM64 or AMD64 automatically. No Swift sources or compiler are needed.

ghcr.io/assistela/assistela:2.0.29

Docker Compose

Save this as compose.yaml, then run the commands below in that folder. This example is accessible only from the Docker host.

Download compose.yaml

name: assistela-docker-v2
services:
  assistela:
    image: ghcr.io/assistela/assistela:2.0.29
    hostname: assistela-docker-v2
    init: true
    restart: unless-stopped
    stop_grace_period: 120s
    shm_size: 512mb
    ports:
      - "127.0.0.1:8234:8234"
    environment:
      ASSISTELA_PUBLIC_URL: http://localhost:8234
      ASSISTELA_TIME_ZONE: Europe/Prague
    volumes:
      - assistela-data:/data
    security_opt:
      - no-new-privileges:true
volumes:
  assistela-data:

docker compose pull
docker compose up -d
docker compose logs assistela

http://localhost:8234/admin/

After installation, open the admin interface on your Docker host, use the setup token from the container logs and choose an admin password. Then connect your AI model and pair the iOS Client with a QR code.

Synology Container Manager

On a compatible Synology NAS, open Container Manager → Project → Create. Choose a folder, paste the YAML below into the editor and start the project.

Replace NAS_IP in both places with the NAS private IPv4 address, for example 192.168.1.10. Open http://NAS_IP:8234/admin/ and find the setup token in the container logs.

Recommended Synology configuration, tested with user: "1026:100". Create /volume1/docker/assistela first and set user to the folder owner’s UID:GID; check your account IDs with id over SSH. Adjust the folder path if needed.

Download Synology compose.yaml

name: assistela
services:
  assistela:
    image: ghcr.io/assistela/assistela:2.0.29
    user: "1026:100"
    hostname: assistela
    init: true
    restart: unless-stopped
    stop_grace_period: 120s
    shm_size: 512mb
    ports:
      - "NAS_IP:8234:8234"
    environment:
      ASSISTELA_PUBLIC_URL: http://NAS_IP:8234
      ASSISTELA_ALLOW_LAN_HTTP: "1"
      ASSISTELA_TIME_ZONE: Europe/Prague
    volumes:
      - /volume1/docker/assistela:/data
    security_opt:
      - no-new-privileges:true

The Synology example uses unencrypted HTTP on your trusted LAN. For internet access, use an HTTPS reverse proxy and set ASSISTELA_PUBLIC_URL to its exact address.

Synology — Container Manager / Project

Updates and backups

Data stays in the storage mounted at /data: the Docker volume or the NAS folder. Keep the storage mapping, project name and hostname unchanged. Back up /data before updating; never delete its contents.

To update, change the image tag to the new version and run pull and up -d again. In Container Manager, edit the YAML and redeploy the project.