slices.network
Auto-indexing service and GraphQL API for AT Protocol Records
quickslice.slices.network/
atproto
gleam
graphql
chadtmiller.com
1# Deployment
2
3Deploy Quickslice to production. Railway with one-click deploy is fastest.
4
5## Railway (Recommended)
6
7### 1. Deploy
8
9Click the button to deploy Quickslice with SQLite:
10
11[](https://railway.com/deploy/quickslice?referralCode=Ofii6e&utm_medium=integration&utm_source=template&utm_campaign=generic)
12
13Railway prompts you to configure environment variables. Leave the form open while you generate a signing key.
14
15### 2. Generate OAuth Signing Key
16
17Quickslice needs a private key to sign OAuth tokens:
18
19```bash
20brew install goat
21goat key generate -t p256
22```
23
24This outputs:
25
26```
27Key Type: P-256 / secp256r1 / ES256 private key
28Secret Key (Multibase Syntax): z42tsQ4W...
29Public Key (DID Key Syntax): did:key:zDnaek...
30```
31
32Copy only the **Secret Key** value (starts with `z`) and paste it into the `OAUTH_SIGNING_KEY` field in Railway, then click **Save Config**.
33
34### 3. Configure Your Domain
35
36After deployment completes:
37
381. Click on your quickslice service
392. Go to **Settings**
403. Click **Generate Domain** under Networking
41
42Railway creates a public URL like `quickslice-production-xxxx.up.railway.app`.
43
44**Redeploy to apply the domain:**
45
461. Go to **Deployments**
472. Click the three-dot menu on the latest deployment
483. Select **Redeploy**
49
50### 4. Create Admin Account
51
52Visit your domain. The welcome screen prompts you to create an admin account:
53
541. Enter your AT Protocol handle (e.g., `yourname.bsky.social`)
552. Click **Authenticate**
563. Authorize Quickslice on your PDS
574. You're now the instance admin
58
59### 5. Configure Your Instance
60
61From the homepage, go to **Settings**:
62
631. Enter your **Domain Authority** in reverse-domain format (e.g., `xyz.statusphere`)
642. Upload your Lexicons as a `.zip` file (JSON format, directory structure doesn't matter). See [statusphere lexicons](https://tangled.org/slices.network/lexicon-sets/tree/main/statusphere) for an example.
65 ```
66 lexicons.zip
67 └── lexicons/
68 └── xyz/
69 └── statusphere/
70 └── status.json
71 ```
723. Click **Trigger Backfill** to import existing records from the network. The Quickslice logo enters a loading state during backfill and the page refreshes when complete. Check Railway logs to monitor progress:
73 ```
74 INFO [backfill] PDS worker 67/87 done (1898 records)
75 INFO [backfill] PDS worker 68/87 done (1117 records)
76 INFO [backfill] PDS worker 69/87 done (746 records)
77 ...
78 ```
79 Depending on the lexicon, this could take a few seconds (`xyz.statusphere.*`) to days (`app.bsky.*`) to complete. Be mindful of your available storage and associated cloud provider fees when backfilling large lexicons.
80
81## Environment Variables
82
83| Variable | Required | Default | Description |
84|----------|----------|---------|-------------|
85| `OAUTH_SIGNING_KEY` | Yes | - | P-256 private key for signing OAuth tokens |
86| `DATABASE_URL` | No | `quickslice.db` | Path to SQLite database |
87| `HOST` | No | `127.0.0.1` | Server bind address (use `0.0.0.0` for containers) |
88| `PORT` | No | `8080` | Server port |
89| `SECRET_KEY_BASE` | Recommended | Auto-generated | Session encryption key (64+ chars) |
90| `EXTERNAL_BASE_URL` | No | Auto-detected | Public URL for OAuth redirects |
91
92## Fly.io
93
94### 1. Create a Volume
95
96```bash
97fly volumes create app_data --size 10
98```
99
100### 2. Configure fly.toml
101
102```toml
103app = 'your-app-name'
104primary_region = 'sjc'
105
106[build]
107 dockerfile = "Dockerfile"
108
109[env]
110 DATABASE_URL = 'sqlite:/data/quickslice.db'
111 EXTERNAL_BASE_URL=https://your-quickslice.fly.dev
112 HOST = '0.0.0.0'
113 PORT = '8080'
114
115[http_service]
116 internal_port = 8080
117 force_https = true
118 auto_stop_machines = 'stop'
119 auto_start_machines = true
120 min_machines_running = 1
121
122[[mounts]]
123 source = 'app_data'
124 destination = '/data'
125
126[[vm]]
127 memory = '1gb'
128 cpu_kind = 'shared'
129 cpus = 1
130```
131
132### 3. Set Secrets
133
134```bash
135fly secrets set SECRET_KEY_BASE=$(openssl rand -base64 48)
136```
137
138Generate a signing key and copy only the **Secret Key** value (starts with `z`):
139
140```bash
141goat key generate -t p256
142fly secrets set OAUTH_SIGNING_KEY="z42tsQ4W..." # paste your secret key here
143```
144
145### 4. Deploy
146
147```bash
148fly deploy
149```
150
151## Docker Compose
152
153For self-hosted deployments:
154
155```yaml
156version: "3.8"
157
158services:
159 quickslice:
160 image: ghcr.io/bigmoves/quickslice:latest
161 ports:
162 - "8080:8080"
163 volumes:
164 - quickslice-data:/data
165 environment:
166 - HOST=0.0.0.0
167 - PORT=8080
168 - DATABASE_URL=sqlite:/data/quickslice.db
169 - SECRET_KEY_BASE=${SECRET_KEY_BASE}
170 - OAUTH_SIGNING_KEY=${OAUTH_SIGNING_KEY}
171 - EXTERNAL_BASE_URL=https://your-quickslice.example.com
172 restart: unless-stopped
173
174volumes:
175 quickslice-data:
176```
177
178Create a `.env` file:
179
180```bash
181SECRET_KEY_BASE=$(openssl rand -base64 48)
182OAUTH_SIGNING_KEY=z42tsQ4W... # paste your secret key here
183```
184
185Start:
186
187```bash
188docker compose up -d
189```
190
191## Backfill Configuration
192
193NOTE: These configurations are evolving. If your container runs low on memory or crashes, reduce concurrent workers and requests.
194
195Control memory usage during backfill with these variables:
196
197| Variable | Default | Description |
198|----------|---------|-------------|
199| `BACKFILL_MAX_PDS_WORKERS` | 10 | Max concurrent PDS endpoints |
200| `BACKFILL_PDS_CONCURRENCY` | 4 | Max concurrent repo fetches per PDS |
201| `BACKFILL_MAX_HTTP_CONCURRENT` | 50 | Global HTTP request limit |
202
203**1GB RAM:**
204```
205BACKFILL_MAX_PDS_WORKERS=8
206BACKFILL_PDS_CONCURRENCY=2
207BACKFILL_MAX_HTTP_CONCURRENT=30
208```
209
210**2GB+ RAM:** Use defaults or increase values.
211
212## Resource Requirements
213
214**Minimum:**
215- Memory: 1GB
216- CPU: 1 shared core
217- Disk: 10GB volume
218
219**Recommendations:**
220- Use SSD-backed volumes for SQLite performance
221- Monitor database size and scale volume as needed
222
223## PostgreSQL Deployment
224
225For deployments requiring a full database server, use the PostgreSQL template:
226
227[](https://railway.com/deploy/GRtFyd?referralCode=Ofii6e&utm_medium=integration&utm_source=template&utm_campaign=generic)
228
229This template provisions a PostgreSQL database alongside Quickslice. The `DATABASE_URL` is automatically configured.