# Template catalog (/templates/catalog)
{/* AUTO-GENERATED by scripts/generate-template-docs.mjs from @alternatefutures/backend@0.1.2. Do not edit by hand. */}
Generated on every merge from the platform's template registry (@alternatefutures/backend 0.1.2), filtered
to what the platform currently offers, so it matches the web app and `acc templates list`. See the
[templates overview](/templates) for how deployment from a template works.
Deploy a template from the web app's **Deploy a template** page at
`https://clouds.alternatefutures.ai/deploy`, or from the CLI with its ID:
```bash
acc templates info
acc services create --kind template --template
```
Composite templates (several services deployed together, marked in the table)
deploy from the web app only; the CLI does not support them yet.
## All templates [#all-templates]
| Template | ID | Category | Deploy from | GPU | vCPU | Memory | Storage |
| --------------------------------------------------------------- | ------------------------------- | ----------------------- | -------------- | ---------- | ---- | ------ | ------- |
| [GPU Instance](#gpu-instance) | `gpu-instance` | AI and machine learning | web app or CLI | 1 × NVIDIA | 8 | 32Gi | 10Gi |
| [Clawd Bot (OpenClaw)](#clawd-bot-openclaw) | `openclaw-gateway` | AI and machine learning | web app or CLI | no | 1 | 2Gi | 2Gi |
| [Ollama GPU](#ollama-gpu) | `ollama-gpu` | AI and machine learning | web app or CLI | 1 × NVIDIA | 4 | 16Gi | 10Gi |
| [Jupyter ML Workspace](#jupyter-ml-workspace) | `jupyter-ml-workspace` | AI and machine learning | web app or CLI | 1 × NVIDIA | 8 | 32Gi | 10Gi |
| [Alternate Agent](#alternate-agent) | `alternate-agent` | AI and machine learning | web app | no | 1 | 2Gi | 2Gi |
| [nanobot](#nanobot) | `nanobot-gateway` | AI and machine learning | web app or CLI | no | 0.5 | 1Gi | 2Gi |
| [ComfyUI](#comfyui) | `comfyui` | AI and machine learning | web app or CLI | 1 × NVIDIA | 4 | 16Gi | 20Gi |
| [Attested Server](#attested-server) | `attested-server` | Game servers | web app or CLI | no | 1 | 2Gi | 10Gi |
| [Node.js Game Server](#node-js-game-server) | `node-ws-gameserver` | Game servers | web app or CLI | no | 0.5 | 512Mi | 1Gi |
| [Bun Game Server](#bun-game-server) | `bun-ws-gameserver` | Game servers | web app or CLI | no | 0.5 | 512Mi | 1Gi |
| [Minecraft Server](#minecraft-server) | `minecraft-server` | Game servers | web app or CLI | no | 2 | 4Gi | 10Gi |
| [Alternate Chat](#alternate-chat) | `alternate-chat` | Web servers | web app or CLI | no | 0.5 | 512Mi | 1Gi |
| [Hello Future](#hello-future) | `hello-future` | Web servers | web app or CLI | no | 0.1 | 64Mi | 256Mi |
| [React + Vite](#react-vite) | `react-vite` | Web servers | web app or CLI | no | 0.25 | 256Mi | 1Gi |
| [Hugo](#hugo) | `hugo-server` | Web servers | web app or CLI | no | 0.25 | 128Mi | 512Mi |
| [PostgreSQL](#postgresql) | `postgres` | Databases | web app or CLI | no | 0.5 | 1Gi | 1Gi |
| [Redis](#redis) | `redis` | Databases | web app or CLI | no | 0.25 | 256Mi | 1Gi |
| [RustFS](#rustfs) | `rustfs` | Storage | web app or CLI | no | 0.5 | 1Gi | 1Gi |
| [Infisical](#infisical) | `infisical` | Developer tools | web app or CLI | no | 1 | 1Gi | 512Mi |
| [n8n](#n8n) | `n8n` | Developer tools | web app or CLI | no | 0.5 | 1Gi | 2Gi |
| [Gitea](#gitea) | `gitea` | Developer tools | web app or CLI | no | 0.5 | 512Mi | 2Gi |
| [Telegram Community Bot Runner](#telegram-community-bot-runner) | `telegram-community-bot-runner` | Developer tools | web app or CLI | no | 0.5 | 1Gi | 4Gi |
## AI and machine learning [#ai-and-machine-learning]
### GPU Instance [#gpu-instance]
Linux machine with NVIDIA GPU and CUDA. Install anything: Ollama, vLLM, training frameworks, or your own stack. Connect via the web terminal or CLI.
ID `gpu-instance` · featured · tags `gpu`, `cuda`, `instance`, `vm`, `linux`, `machine-learning`, `training`
* **Deploy**: `acc services create --kind template --template gpu-instance` or from the web app.
* **Image**: `nvidia/cuda:12.8.1-devel-ubuntu24.04` · [source](https://github.com/NVIDIA/nvidia-docker)
* **Resources**: 8 vCPU, 32Gi memory, 10Gi storage, GPU 1 × NVIDIA
* **Ports**: container 8080 as 80 (public)
* **Persistent storage**: `home` 100Gi at `/root/workspace`
No environment variables to configure.
### Clawd Bot (OpenClaw) [#clawd-bot-openclaw]
Personal AI assistant gateway (OpenClaw/Clawd Bot) with Control UI, multi-provider models, channels, and optional sandboxed tools.
ID `openclaw-gateway` · featured · tags `ai`, `assistant`, `agent`, `openclaw`, `clawd`, `gateway`
* **Deploy**: `acc services create --kind template --template openclaw-gateway` or from the web app.
* [source](https://github.com/openclaw/openclaw)
* **Resources**: 1 vCPU, 2Gi memory, 2Gi storage
* **Ports**: container 18789 as 80 (public); container 18790 as 18790 (public)
* **Persistent storage**: `openclaw-state` 10Gi at `/home/node/.openclaw`
| Variable | Required | Default | Description |
| ------------------------------ | -------- | ---------------------- | -------------------------------------------------------------------------------- |
| `OPENCLAW_GATEWAY_TOKEN` | yes | none (secret) | Control UI token (recommended): used to authenticate dashboard/control UI access |
| `OPENCLAW_GATEWAY_PASSWORD` | no | none (secret) | Alternative to token auth: password for Control UI access |
| `ANTHROPIC_API_KEY` | no | none (secret) | Anthropic API key for Claude models (recommended) |
| `OPENAI_API_KEY` | no | none (secret) | OpenAI API key (optional provider) |
| `GOOGLE_GENERATIVE_AI_API_KEY` | no | none (secret) | Google Gemini API key (optional provider) |
| `CLAUDE_AI_SESSION_KEY` | no | none (secret) | Claude subscription session key (advanced; optional) |
| `CLAUDE_WEB_SESSION_KEY` | no | none (secret) | Claude web session key (advanced; optional) |
| `CLAUDE_WEB_COOKIE` | no | none (secret) | Claude web cookie (advanced; optional) |
| `OPENCLAW_STATE_DIR` | yes | `/home/node/.openclaw` | Directory for OpenClaw config, agent state, and workspace |
### Ollama GPU [#ollama-gpu]
Run any open LLM (Llama 70B, Qwen 72B, Mistral, etc.) on GPU. OpenAI-compatible API included. Pull models on demand.
ID `ollama-gpu` · featured · tags `ai`, `llm`, `inference`, `gpu`, `ollama`, `openai`
* **Deploy**: `acc services create --kind template --template ollama-gpu` or from the web app.
* **Image**: `ollama/ollama:0.30.7` · [source](https://github.com/ollama/ollama)
* **Resources**: 4 vCPU, 16Gi memory, 10Gi storage, GPU 1 × NVIDIA
* **Ports**: container 11434 as 80 (public)
* **Persistent storage**: `ollama-models` 50Gi at `/data/models`
| Variable | Required | Default | Description |
| --------------- | -------- | -------------- | --------------------------------------------------- |
| `OLLAMA_HOST` | yes | `0.0.0.0` | Listen address (0.0.0.0 for all interfaces) |
| `OLLAMA_MODELS` | yes | `/data/models` | Directory for downloaded model weights (persistent) |
### Jupyter ML Workspace [#jupyter-ml-workspace]
GPU-powered Jupyter Lab with PyTorch and CUDA pre-installed. Fine-tune models, run experiments, and train on cloud GPU.
ID `jupyter-ml-workspace` · featured · tags `ai`, `ml`, `jupyter`, `pytorch`, `gpu`, `training`, `fine-tuning`
* **Deploy**: `acc services create --kind template --template jupyter-ml-workspace` or from the web app.
* **Image**: `quay.io/jupyter/pytorch-notebook:cuda12-pytorch-2.5.1` · [source](https://github.com/jupyter/docker-stacks)
* **Resources**: 8 vCPU, 32Gi memory, 10Gi storage, GPU 1 × NVIDIA
* **Ports**: container 8888 as 80 (public)
* **Persistent storage**: `jupyter-work` 100Gi at `/home/jovyan/work`
| Variable | Required | Default | Description |
| --------------- | -------- | ------------- | ----------------------------------------------- |
| `JUPYTER_TOKEN` | yes | none (secret) | Access token for Jupyter Lab (acts as password) |
### Alternate Agent [#alternate-agent]
AI agent with chat UI, persistent memory, autonomous goal loop, and pluggable connectors (Discord, X/Twitter). Native Supabase RAG + image/video generation through AF inference. No separate API keys: uses your AF wallet.
ID `alternate-agent` · featured · tags `ai`, `agent`, `mastra`, `autonomous`, `rag`, `supabase`, `x`, `twitter`, `social`, `chatbot`, `discord`
* **Deploy**: from the web app only. Composite template (`PostgreSQL + pgvector`, `Alternate Agent`); the CLI does not support composite templates yet.
* **Image**: `ghcr.io/alternatefutures/alternate-agent:v2` · [source](https://github.com/alternatefutures/alternate-agent)
* **Resources**: 1 vCPU, 2Gi memory, 2Gi storage
* **Ports**: container 3000 as 80 (public)
* **Persistent storage**: `agent-data` 5Gi at `/app/data`
* **Health check**: `/health` on port 3000
| Variable | Required | Default | Description |
| ------------------------- | -------- | --------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| `AF_API_KEY` | no | set by the platform (`apiKey`) | Alternate Clouds PAT: funds chat, image, and video inference. |
| `AF_ORG_ID` | no | set by the platform (`orgId`) | Alternate Clouds organization ID (for billing). |
| `MODEL_NAME` | no | `gpt-4o-mini` | Chat model (e.g. gpt-4o-mini, gpt-4o, claude-sonnet-4, grok-3). |
| `AGENT_NAME` | no | `Alternate Agent` | Agent display name shown in chat UI and bot identity. |
| `AGENT_INSTRUCTIONS` | no | `You are a helpful AI assistant.` | System prompt / personality. Use this to set tone, voice, constraints. |
| `AGENT_GOAL` | no | none | Sets the agent to autonomous mode. Free-form description of what the agent should pursue every tick (e.g. "Grow @myproject on X by sharing project updates and replying to mentions, grounded in the project KB"). Leave blank for chat-only. |
| `AUTONOMOUS_INTERVAL_MIN` | no | `30` | Minutes between autonomous ticks (5-1440). ±20% jitter applied. |
| `SUPABASE_URL` | no | none | Your Supabase project URL (e.g. [https://xxx.supabase.co](https://xxx.supabase.co)). Enables the query\_knowledge\_base tool for RAG over your existing vector store. |
| `SUPABASE_KEY` | no | none (secret) | Supabase service role or anon key. Service role recommended so the agent can call the RPC even if you have RLS enabled. |
| `SUPABASE_RPC_NAME` | no | `match_documents` | Name of the Supabase RPC for vector similarity search. Default matches the standard Supabase pgvector quickstart. Function signature: (query text, match\_count int) returns table(content text, similarity float). |
| `TWITTER_COOKIES` | no | none (secret) | X session cookies as JSON (recommended). Capture from a logged-in browser session. Survives X anti-automation far better than password login. Format: array of cookie strings or cookie objects. |
| `TWITTER_USERNAME` | no | none | X handle (without @). Fallback if TWITTER\_COOKIES is not set. |
| `TWITTER_PASSWORD` | no | none (secret) | X password. Triggers password login (fragile; X often challenges with captcha). |
| `TWITTER_EMAIL` | no | none | X account email. Sometimes required by the password login flow. |
| `DISCORD_BOT_TOKEN` | no | none (secret) | Discord bot token: enables the Discord connector for chat in servers. |
| `IMAGE_GEN_MODEL` | no | `fal-ai/flux/schnell` | fal.ai image model path. Default is cheap & fast (\~$0.003/image). Try fal-ai/flux/dev or fal-ai/imagen3 for higher quality. |
| `VIDEO_GEN_MODEL` | no | `fal-ai/veo3/fast` | fal.ai video model path. EXPENSIVE per call (\~$0.50+). Set conservative maxBudgetUsd in deployment policy to cap blast radius. |
| `AUTH_TOKEN` | no | none (secret) | Optional Bearer token to protect the HTTP API at /api/\*. |
### nanobot [#nanobot]
Ultra-lightweight personal AI assistant (\~4k lines). Multi-provider (OpenRouter, Claude, GPT), CLI + chat channels (Telegram, Discord, Slack). No web UI.
ID `nanobot-gateway` · tags `ai`, `assistant`, `agent`, `nanobot`, `openclaw`, `gateway`, `lightweight`
* **Deploy**: `acc services create --kind template --template nanobot-gateway` or from the web app.
* [source](https://github.com/HKUDS/nanobot)
* **Resources**: 0.5 vCPU, 1Gi memory, 2Gi storage
* **Ports**: container 18790 as 80 (public)
* **Persistent storage**: `nanobot-state` 10Gi at `/home/nanobot/.nanobot`
| Variable | Required | Default | Description |
| ----------------------- | -------- | --------------------------- | ------------------------------------------------------------------------------------------------------------- |
| `OPENROUTER_API_KEY` | yes | none (secret) | OpenRouter API key (recommended: access to all models). Required: nanobot exits without at least one LLM key. |
| `ANTHROPIC_API_KEY` | no | none (secret) | Anthropic API key for Claude models |
| `OPENAI_API_KEY` | no | none (secret) | OpenAI API key for GPT models |
| `DEEPSEEK_API_KEY` | no | none (secret) | DeepSeek API key (optional) |
| `GROQ_API_KEY` | no | none (secret) | Groq API key (optional: also enables Whisper voice transcription) |
| `NANOBOT_DEFAULT_MODEL` | yes | `anthropic/claude-opus-4-5` | Default LLM model (e.g. anthropic/claude-opus-4-5, openrouter/...) |
| `NANOBOT_STATE_DIR` | yes | `/home/nanobot/.nanobot` | Directory for config, workspace, and session data |
### ComfyUI [#comfyui]
Node-based Stable Diffusion UI with GPU acceleration. Build complex image generation pipelines with a visual workflow editor. Supports SD 1.5, SDXL, and Flux.
ID `comfyui` · tags `ai`, `stable-diffusion`, `image-generation`, `gpu`, `comfyui`, `generative`
* **Deploy**: `acc services create --kind template --template comfyui` or from the web app.
* **Image**: `yanwk/comfyui-boot:cu128-slim` · [source](https://github.com/comfyanonymous/ComfyUI)
* **Resources**: 4 vCPU, 16Gi memory, 20Gi storage, GPU 1 × NVIDIA
* **Ports**: container 8188 as 80 (public)
* **Persistent storage**: `comfyui-models` 50Gi at `/root/ComfyUI/models`; `comfyui-output` 10Gi at `/root/ComfyUI/output`
| Variable | Required | Default | Description |
| ---------- | -------- | ------------------------------ | ------------------------------ |
| `CLI_ARGS` | yes | `--listen 0.0.0.0 --port 8188` | ComfyUI command-line arguments |
## Game servers [#game-servers]
### Attested Server [#attested-server]
Your game/app image in a hard-locked Intel TDX enclave with a hardware-attested identity. No shell, no file access: nobody (not even you) can tamper with the box, and anyone can verify what it runs against Intel roots. Set your own Docker image (listening on port 80) before deploying; env vars pass through. Identity is served at the \-8511 gateway domain for registry registration. Confidential compute only.
ID `attested-server` · tags `tee`, `attested`, `tdx`, `confidential`, `game-server`, `provenance`, `locked`
* **Deploy**: `acc services create --kind template --template attested-server` or from the web app.
* [source](https://github.com/alternatefutures/alternate-clouds-api/tree/main/docker/attested-identity)
* **Resources**: 1 vCPU, 2Gi memory, 10Gi storage
* **Ports**: container 80 as 80 (public)
* **Capabilities**: attested identity (confidential compute with a verifiable TDX quote); no interactive shell, by design
No environment variables to configure.
### Node.js Game Server [#nodejs-game-server]
Protocol-agnostic WebSocket relay server with room-based architecture and binary protocol (msgpack). Relays any game data between peers without inspecting payloads.
ID `node-ws-gameserver` · tags `websocket`, `multiplayer`, `realtime`, `gameserver`, `msgpack`, `node`
* **Deploy**: `acc services create --kind template --template node-ws-gameserver` or from the web app.
* **Image**: `ghcr.io/mavisakalyan/node-ws-gameserver:046edbd` · [source](https://github.com/mavisakalyan/node-ws-gameserver)
* **Resources**: 0.5 vCPU, 512Mi memory, 1Gi storage
* **Ports**: container 8080 as 80 (public)
* **Health check**: `/health` on port 8080
| Variable | Required | Default | Description |
| ------------------------- | -------- | ------- | --------------------------------------------------------- |
| `PORT` | yes | `8080` | Server listen port |
| `ALLOWED_ORIGINS` | no | `*` | Comma-separated allowed origins for WebSocket connections |
| `KEEPALIVE_MS` | no | `30000` | WebSocket keepalive ping interval in milliseconds |
| `MAX_MESSAGES_PER_SECOND` | no | `60` | Per-client rate limit (sliding window) |
| `MAX_PLAYERS_PER_ROOM` | no | `50` | Maximum players per room |
### Bun Game Server [#bun-game-server]
Bun-native WebSocket relay server: same protocol as the Node.js version but 5-8x faster. Protocol-agnostic relay with Bun built-in pub/sub for room broadcasting.
ID `bun-ws-gameserver` · tags `websocket`, `multiplayer`, `realtime`, `gameserver`, `msgpack`, `bun`
* **Deploy**: `acc services create --kind template --template bun-ws-gameserver` or from the web app.
* **Image**: `ghcr.io/mavisakalyan/bun-ws-gameserver:7ad7059` · [source](https://github.com/mavisakalyan/bun-ws-gameserver)
* **Resources**: 0.5 vCPU, 512Mi memory, 1Gi storage
* **Ports**: container 8080 as 80 (public)
* **Health check**: `/health` on port 8080
| Variable | Required | Default | Description |
| ------------------------- | -------- | ------- | --------------------------------------------------------- |
| `PORT` | yes | `8080` | Server listen port |
| `ALLOWED_ORIGINS` | no | `*` | Comma-separated allowed origins for WebSocket connections |
| `KEEPALIVE_MS` | no | `30000` | WebSocket keepalive ping interval in milliseconds |
| `MAX_MESSAGES_PER_SECOND` | no | `60` | Per-client rate limit (sliding window) |
| `MAX_PLAYERS_PER_ROOM` | no | `50` | Maximum players per room |
### Minecraft Server [#minecraft-server]
Minecraft Java Edition server with itzg's production-hardened image. Supports Vanilla, Paper, Fabric, Forge, and more. Auto-downloads server JAR on first start.
ID `minecraft-server` · tags `minecraft`, `gameserver`, `java`, `multiplayer`, `game`
* **Deploy**: `acc services create --kind template --template minecraft-server` or from the web app.
* **Image**: `itzg/minecraft-server:java21` · [source](https://github.com/itzg/docker-minecraft-server)
* **Resources**: 2 vCPU, 4Gi memory, 10Gi storage
* **Ports**: container 25565 as 25565 (public); container 25575 as 25575 (internal)
* **Persistent storage**: `minecraft-data` 20Gi at `/data`
| Variable | Required | Default | Description |
| --------------- | -------- | ---------------------------------------- | ---------------------------------------------------------- |
| `EULA` | yes | `TRUE` | Accept Minecraft EULA (must be TRUE) |
| `TYPE` | yes | `PAPER` | Server type: VANILLA, PAPER, FABRIC, FORGE, BUKKIT, SPIGOT |
| `VERSION` | yes | `LATEST` | Minecraft version (e.g. 1.21.4 or LATEST) |
| `MEMORY` | yes | `2G` | JVM heap size (e.g. 2G, 4G) |
| `DIFFICULTY` | no | `normal` | Game difficulty: peaceful, easy, normal, hard |
| `MAX_PLAYERS` | no | `20` | Maximum number of players |
| `MOTD` | no | `A Minecraft Server on Alternate Clouds` | Message of the day shown in the server list |
| `RCON_PASSWORD` | no | none (secret) | RCON remote console password |
## Web servers [#web-servers]
### Alternate Chat [#alternate-chat]
End-to-end encrypted group chat where the passphrase is the room. Messages are sealed in your browser (AES-256-GCM, Argon2id keys, Ed25519-signed): the relay node only ever sees ciphertext, so no ISP, node operator, or AlternateFutures can read them. You own the deployment.
ID `alternate-chat` · featured · tags `AES-256-GCM`, `Argon2id`, `Ed25519-signed`, `chat`, `e2ee`, `encryption`, `privacy`, `websocket`, `realtime`
* **Deploy**: `acc services create --kind template --template alternate-chat` or from the web app.
* **Image**: `ghcr.io/alternatefutures/alternate-chat:v102` · [source](https://github.com/alternatefutures/alternate-chat)
* **Resources**: 0.5 vCPU, 512Mi memory, 1Gi storage
* **Ports**: container 8080 as 80 (public)
* **Persistent storage**: `data` 1Gi at `/data`
* **Health check**: `/healthz` on port 8080
| Variable | Required | Default | Description |
| ------------------------- | -------- | ------- | --------------------------------------------------------------------------- |
| `PORT` | yes | `8080` | HTTP/WebSocket listen port (served behind the AF edge proxy). |
| `DATA_DIR` | yes | `/data` | Directory for the SQLite database (must match the persistent volume mount). |
| `HISTORY_LIMIT` | no | `200` | Most recent messages replayed to a client when it joins a room. |
| `MAX_MESSAGES_PER_SECOND` | no | `20` | Per-connection publish rate limit (sliding 1s window). |
| `MAX_CIPHERTEXT_BYTES` | no | `65536` | Maximum accepted ciphertext size per message (bytes). |
### Hello Future [#hello-future]
A minimal static HTML page served by Nginx. Defaults to a blue "Hello Future" page; pass HTML\_CONTENT (base64-encoded HTML) to deploy your own page.
ID `hello-future` · tags `web`, `static`, `nginx`, `hello-future`, `demo`
* **Deploy**: `acc services create --kind template --template hello-future` or from the web app.
* **Image**: `nginx:1.27-alpine` · [source](https://github.com/nginx/nginx)
* **Resources**: 0.1 vCPU, 64Mi memory, 256Mi storage
* **Ports**: container 80 as 80 (public)
* **Health check**: `/` on port 80
| Variable | Required | Default | Description |
| -------------- | -------- | ------------------------------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ |
| `HTML_CONTENT` | no | `PCFET0NUWVBFIGh0bWw+CjxodG1sPgo8aGVhZD4KPG1ld…` | Base64-encoded HTML to serve as index.html. Use `base64 -w 0 file.html` (Linux) or `base64 -i file.html \| tr -d "\n"` (macOS): the value MUST NOT contain newlines (the GPU compute tier .env files reject them). |
### React + Vite [#react--vite]
React SPA served by Nginx. Built with Vite for fast builds and hot module replacement. Perfect for client-side rendered applications.
ID `react-vite` · tags `web`, `react`, `vite`, `spa`, `nginx`, `frontend`
* **Deploy**: `acc services create --kind template --template react-vite` or from the web app.
* **Image**: `nginx:1.27-alpine` · [source](https://github.com/vitejs/vite)
* **Resources**: 0.25 vCPU, 256Mi memory, 1Gi storage
* **Ports**: container 80 as 80 (public)
* **Health check**: `/` on port 80
| Variable | Required | Default | Description |
| ------------ | -------- | ------- | ----------------- |
| `NGINX_PORT` | yes | `80` | Nginx listen port |
### Hugo [#hugo]
Hugo static site generator served by Nginx. One of the fastest static site generators, with Markdown-driven content and theme support.
ID `hugo-server` · tags `web`, `hugo`, `static`, `blog`, `nginx`, `go`
* **Deploy**: `acc services create --kind template --template hugo-server` or from the web app.
* **Image**: `nginx:1.27-alpine` · [source](https://github.com/gohugoio/hugo)
* **Resources**: 0.25 vCPU, 128Mi memory, 512Mi storage
* **Ports**: container 80 as 80 (public)
* **Health check**: `/` on port 80
No environment variables to configure.
## Databases [#databases]
### PostgreSQL [#postgresql]
PostgreSQL 17 Alpine: lightweight, production-ready relational database with persistent storage.
ID `postgres` · tags `database`, `sql`, `postgres`, `relational`
* **Deploy**: `acc services create --kind template --template postgres` or from the web app.
* **Image**: `postgres:17-alpine` · [source](https://hub.docker.com/_/postgres)
* **Resources**: 0.5 vCPU, 1Gi memory, 1Gi storage
* **Ports**: container 5432 as 5432 (public)
* **Persistent storage**: `pgdata` 2Gi at `/var/lib/postgresql/data`
* **Capabilities**: exposes connection strings `DATABASE_URL`, `PGHOST`, `PGPORT`, `PGUSER`, `PGPASSWORD`, `PGDATABASE` to linked services
| Variable | Required | Default | Description |
| ------------------- | -------- | --------------------------------- | --------------------------------------------------------------------- |
| `POSTGRES_DB` | yes | `appdb` | Default database name |
| `POSTGRES_USER` | yes | `postgres` | Database superuser name |
| `POSTGRES_PASSWORD` | yes | none (secret) | Database superuser password |
| `PGDATA` | yes | `/var/lib/postgresql/data/pgdata` | PostgreSQL data directory (must be a subdirectory of the mount point) |
### Redis [#redis]
Redis 7 Alpine: in-memory data store for caching, sessions, pub/sub, and rate limiting. Optional persistent storage.
ID `redis` · tags `database`, `cache`, `redis`, `nosql`, `pubsub`
* **Deploy**: `acc services create --kind template --template redis` or from the web app.
* **Image**: `redis:7-alpine` · [source](https://hub.docker.com/_/redis)
* **Resources**: 0.25 vCPU, 256Mi memory, 1Gi storage
* **Ports**: container 6379 as 6379 (public)
* **Persistent storage**: `redisdata` 5Gi at `/data`
* **Capabilities**: exposes connection strings `REDIS_URL`, `REDIS_HOST`, `REDIS_PORT` to linked services
| Variable | Required | Default | Description |
| ------------ | -------- | -------------------------------- | --------------------------------- |
| `REDIS_ARGS` | no | `--save 60 1 --loglevel warning` | Additional Redis server arguments |
## Storage [#storage]
### RustFS [#rustfs]
Self-hosted, S3-compatible object storage powered by RustFS, with persistent volumes and a web console. Drops in anywhere an S3 SDK or CLI works.
ID `rustfs` · featured · tags `storage`, `s3`, `object-storage`, `bucket`, `self-hosted`, `decentralized`, `r2-compatible`, `rustfs`
* **Deploy**: `acc services create --kind template --template rustfs` or from the web app.
* **Image**: `ghcr.io/alternatefutures/rustfs:1.0.0-beta.1` · [source](https://github.com/rustfs/rustfs)
* **Resources**: 0.5 vCPU, 1Gi memory, 1Gi storage
* **Ports**: container 9001 as 80 (public); container 9000 as 9000 (public)
* **Persistent storage**: `data` 10Gi at `/data`
* **Capabilities**: exposes connection strings `AWS_ENDPOINT_URL_S3`, `S3_ENDPOINT`, `AWS_ACCESS_KEY_ID`, `AWS_SECRET_ACCESS_KEY`, `AWS_REGION`, `AWS_S3_FORCE_PATH_STYLE` to linked services
| Variable | Required | Default | Description |
| ------------------------------------- | -------- | ------------------------------------------ | ------------------------------------------------------------------------------------------------------------------- |
| `RUSTFS_ACCESS_KEY` | no | set by the platform (`generatedAccessKey`) | Root S3 access key ID (auto-generated at deploy time: visible in the deploy form, override to BYO credentials). |
| `RUSTFS_SECRET_KEY` | no | set by the platform (`generatedSecret`) | Root S3 secret access key (auto-generated at deploy time: visible in the deploy form, override to BYO credentials). |
| `RUSTFS_OBS_LOGGER_LEVEL` | no | `info` | Logger verbosity: error \| warn \| info \| debug \| trace |
| `RUSTFS_CORS_ALLOWED_ORIGINS` | no | `*` | CORS allow-list for the S3 API. Use a comma-separated list of origins in production. |
| `RUSTFS_CONSOLE_CORS_ALLOWED_ORIGINS` | no | `*` | CORS allow-list for the web console. Use a comma-separated list of origins in production. |
## Developer tools [#developer-tools]
### Infisical [#infisical]
Self-hosted secrets management: open-source alternative to AWS Secrets Manager / HashiCorp Vault. Stores API keys, DB passwords, certs with end-to-end encryption.
ID `infisical` · tags `secrets`, `security`, `devtools`, `infisical`, `vault`
* **Deploy**: `acc services create --kind template --template infisical` or from the web app.
* **Image**: `infisical/infisical:v0.160.12` · [source](https://github.com/Infisical/infisical)
* **Resources**: 1 vCPU, 1Gi memory, 512Mi storage
* **Ports**: container 8080 as 80 (public)
* **Health check**: `/api/status` on port 8080
| Variable | Required | Default | Description |
| ------------------- | -------- | ----------------------------- | ---------------------------------------------------------------------------------- |
| `ENCRYPTION_KEY` | yes | none (secret) | 32-character hex (16 bytes) for AES-128-GCM. Generate with `openssl rand -hex 16`. |
| `AUTH_SECRET` | yes | none (secret) | 32+ char base64 for JWT signing. Generate with `openssl rand -base64 32`. |
| `SMTP_PASSWORD` | yes | none (secret) | SMTP password (Resend API key recommended, format `re_...`). |
| `SITE_URL` | yes | `https://secrets.example.com` | Public URL of this deployment (used in email links). |
| `SMTP_HOST` | no | `smtp.resend.com` | SMTP server hostname. |
| `SMTP_PORT` | no | `587` | SMTP server port. |
| `SMTP_USERNAME` | no | `resend` | SMTP username. |
| `SMTP_FROM_ADDRESS` | no | `noreply@example.com` | From address for outbound email. |
| `SMTP_FROM_NAME` | no | `Secrets` | From display name for outbound email. |
| `TELEMETRY_ENABLED` | no | `false` | Send anonymous usage telemetry to Infisical. |
### n8n [#n8n]
Workflow automation platform: the open-source alternative to Zapier. Connect 400+ apps with a visual editor, code nodes, and AI-powered automations.
ID `n8n` · tags `automation`, `workflow`, `devtools`, `n8n`, `no-code`, `integration`
* **Deploy**: `acc services create --kind template --template n8n` or from the web app.
* **Image**: `n8nio/n8n:2.6.3` · [source](https://github.com/n8n-io/n8n)
* **Resources**: 0.5 vCPU, 1Gi memory, 2Gi storage
* **Ports**: container 5678 as 80 (public)
* **Persistent storage**: `n8n-data` 10Gi at `/home/node/.n8n`
* **Health check**: `/healthz` on port 5678
| Variable | Required | Default | Description |
| -------------------- | -------- | ------------- | ------------------------------------------------------------------------ |
| `N8N_HOST` | yes | `0.0.0.0` | n8n server listen host |
| `N8N_PORT` | yes | `5678` | n8n server listen port |
| `N8N_PROTOCOL` | yes | `https` | Protocol used for webhooks (http or https) |
| `WEBHOOK_URL` | yes | none | Public URL for incoming webhooks (set to your deployed URL) |
| `N8N_ENCRYPTION_KEY` | yes | none (secret) | Encryption key for stored credentials (generate a random 32-char string) |
| `GENERIC_TIMEZONE` | no | `UTC` | Default timezone for scheduled workflows |
### Gitea [#gitea]
Self-hosted Git service: GitHub/GitLab alternative written in Go. Lightweight, fast, and packed with features: issues, pull requests, CI/CD, packages.
ID `gitea` · tags `git`, `devtools`, `vcs`, `self-hosted`, `go`, `gitea`
* **Deploy**: `acc services create --kind template --template gitea` or from the web app.
* **Image**: `gitea/gitea:1.25-rootless` · [source](https://github.com/go-gitea/gitea)
* **Resources**: 0.5 vCPU, 512Mi memory, 2Gi storage
* **Ports**: container 3000 as 80 (public); container 2222 as 2222 (public)
* **Persistent storage**: `gitea-data` 20Gi at `/var/lib/gitea`
| Variable | Required | Default | Description |
| ----------------------------- | -------- | ----------------------- | --------------------------------------------------------------- |
| `GITEA__server__ROOT_URL` | yes | `http://localhost:3000` | Public URL of the Gitea instance (update to your deployed URL) |
| `GITEA__server__HTTP_PORT` | yes | `3000` | HTTP listen port |
| `GITEA__database__DB_TYPE` | yes | `sqlite3` | Database type (sqlite3, postgres, mysql) |
| `GITEA__security__SECRET_KEY` | yes | none (secret) | Secret key for token signing (generate a random 64-char string) |
### Telegram Community Bot Runner [#telegram-community-bot-runner]
One-click Telegram bot runner for project communities. Add a bot token, project context, links, and guardrails; it runs 24/7 without webhooks.
ID `telegram-community-bot-runner` · tags `telegram`, `community`, `bot`, `faq`, `support`
* **Deploy**: `acc services create --kind template --template telegram-community-bot-runner` or from the web app.
* **Image**: `ghcr.io/alternatefutures/telegram-community-bot:v2`
* **Resources**: 0.5 vCPU, 1Gi memory, 4Gi storage
* **Ports**: container 3000 as 80 (public)
* **Health check**: `/` on port 3000
| Variable | Required | Default | Description |
| ----------------------- | -------- | ------------------------------------------------ | --------------------------------------------------------------- |
| `TELEGRAM_BOT_TOKEN` | yes | none (secret) | Telegram bot token from @BotFather |
| `TELEGRAM_BOT_USERNAME` | no | `telegramchatdemobot` | Telegram bot username (without @) |
| `PROJECT_NAME` | yes | `My Project` | Project or community name the bot represents |
| `PROJECT_CONTEXT` | yes | `Describe what your project does and who it is…` | Source-of-truth project explanation used in bot replies |
| `PROJECT_STATUS` | no | `Early community stage.` | Current status, launch stage, or availability note |
| `PRIMARY_LINK` | no | none | Main link to share with the community |
| `SECONDARY_LINK` | no | none | Optional secondary link such as docs, app, repo, or leaderboard |
| `COMMUNITY_GOAL` | no | none | What the community should do next or understand |
| `BOT_PERSONA` | no | `Be concise, friendly, and factual.` | Tone and style for bot responses |
| `GUARDRAILS` | no | `Do not claim a token has launched unless the …` | Claims the bot must avoid or constraints it must follow |
| `PORT` | yes | `3000` | Health endpoint port used by platform checks |