---
title: "Deploy Synaplan"
description: "Open-source AI platform for your knowledge, models and chat channels"
category: "AI/ML"
url: https://railway.com/deploy/synaplan
---

# Deploy Synaplan

Open-source AI platform for your knowledge, models and chat channels

**[Deploy Synaplan on Railway](https://railway.com/template/synaplan)**

Machine-readable deploy manifest (JSON, validated by TemplateCI): https://railway.com/deploy/synaplan/manifest.json

- **Creator:** My Projects
- **Category:** AI/ML

## Template content

### MariaDB https://github.com/MariaDB.png

- **Image:** mariadb:13.0.2
- **Start command:** `docker-entrypoint.sh --max-allowed-packet=256M --innodb-buffer-pool-size=256M`

### Tika https://github.com/apache.png

- **Image:** apache/tika:3.3.3.0
- **Start command:** `/bin/sh -c 'exec java -cp "/tika-server-standard-${TIKA_VERSION}.jar:/tika-extras/*" org.apache.tika.server.core.TikaServerCli -h "[::]" -p 9998'`

### Synaplan https://github.com/metadist.png

- **Image:** ghcr.io/metadist/synaplan:5.8.9
- **Start command:** `/bin/bash -c 'export OLLAMA_BASE_URL="${OLLAMA_BASE_URL:-}" REALTIME_PUBLIC_WS_URL="${REALTIME_PUBLIC_WS_URL:-}"; H=/tmp/synaplan-helper; mkdir -p "$H"; cp -r /usr/local/bin/lib "$(command -v docker-entrypoint.sh)" "$H/"; sed -i -e "s/^prepare_role_cache() {/prepare_role_cache() { return 0;/" -e "s#php bin/console#php -d max_execution_time=0 bin/console#g" "$H/lib/container-runtime.sh"; run_role() { until php -r "exit(@file_get_contents(\"http://127.0.0.1/api/health\") === false ? 1 : 0);"; do sleep 5; done; while true; do SYNAPLAN_ROLE=$1 SYNAPLAN_WATCHDOG=0 PLUGINS_DIR=/nonexistent "$H/docker-entrypoint.sh" || true; echo "synaplan $1 role exited, restarting in 5s"; sleep 5; done; }; run_role worker & run_role scheduler & exec env SYNAPLAN_ROLE=web docker-entrypoint.sh'`
- **Health check:** /api/health
- **Public domain:** Yes

### Redis https://github.com/redis.png

- **Image:** redis:8.10-alpine
- **Start command:** `/bin/sh -c 'exec redis-server --appendonly yes --appendfsync everysec --maxmemory 256mb --maxmemory-policy allkeys-lru --requirepass "$REDIS_PASSWORD"'`

## Documentation

# Deploy and Host Synaplan on Railway
[Synaplan](https://github.com/metadist/synaplan) is an open-source AI platform for your own knowledge, models and channels: chat with many model providers, upload documents for retrieval-augmented answers, and reach the same assistant from the web, a website widget, email or messengers. This template runs a private Synaplan with its database, cache and document parser.
## About Hosting Synaplan
This template deploys four services from published images:
- **Synaplan** (`ghcr.io/metadist/synaplan`): the web app and API on a public HTTPS domain, plus the background worker and scheduler in the same container, with a volume for uploaded files
- **MariaDB**: the database, including document search on MariaDB's vector type, on a volume
- **Redis**: queue, cache and locks, on a volume
- **Tika**: Apache Tika for text extraction from PDFs and office files, private only
It matches upstream's "lite" quickstart: spoken answers, Qdrant memories and live widget takeover are left out.
**After deploy:**
1. Open the Synaplan service's **Variables** tab and copy `BOOTSTRAP_ADMIN_PASSWORD`.
2. Open https://YOUR-RAILWAY-DOMAIN/login and sign in with `BOOTSTRAP_ADMIN_EMAIL` (default `admin@synaplan.local`) and that password.
3. Add at least one model key (`GROQ_API_KEY`, `OPENAI_API_KEY`, `ANTHROPIC_API_KEY` or `GOOGLE_GEMINI_API_KEY`) to the service variables, or in the app. Synaplan's free Demo tier is switched on too, but its server may refuse requests from cloud hosts.
Self-sign-up and guest chat are off; invite users from the admin area.
## Common Use Cases
- A private ChatGPT-style assistant for a team, with the model provider of your choice
- Answers grounded in your own PDFs and documents
- One assistant reachable from the web, a website chat widget and email
- Trying many models side by side without sending data to a hosted SaaS
## Dependencies for Synaplan Hosting
- An API key for Groq, OpenAI, Anthropic or Google Gemini (or an Ollama server)
- Optional SMTP settings (`MAILER_DSN`) for password resets and email channels
### Deployment Dependencies
- Synaplan: https://github.com/metadist/synaplan (Apache-2.0)
- Configuration reference: https://github.com/metadist/synaplan/blob/main/docs/CONFIGURATION.md
### Implementation Details
- Upstream runs web, worker and scheduler as three containers that share the uploads folder. A Railway volume belongs to one service, so the start command runs the worker and scheduler next to the web role in one container, once the web role has applied migrations and answers its health check. Each restarts on its own if it exits.
- `APP_SECRET`, `TOKEN_SECRET`, the database and Redis passwords and the first admin password are generated at deploy time.
- `APP_URL` and `FRONTEND_URL` follow the Railway domain. If you add a custom domain, update both.
- Pin a newer release by changing the image tag and `APP_VERSION` together; migrations run on start.
## Why Deploy Synaplan on Railway?
Railway is a singular platform to deploy your infrastructure stack. Railway will host your infrastructure so you don't have to deal with configuration, while allowing you to vertically and horizontally scale it.
By deploying Synaplan on Railway, you are one step closer to supporting a complete full-stack application with minimal burden. Host your servers, databases, AI agents, and more on Railway.

## Similar templates

- [Chat Chat](https://railway.com/deploy/-WWW5r) — Chat Chat, your own unified chat and search to AI platform.
- [Weaviate](https://railway.com/deploy/weaviate-1) — Vector database with hybrid search and an API key-protected API
- [Möbius Personal AI OS](https://railway.com/deploy/mobius-v01-personal-ai-os) — Self-hosted AI agent that builds and runs apps for you, with memory.

Open this page in a browser: https://railway.com/deploy/synaplan
