Skip to content

Execution Guide — Operating the Dev Station

Daily operations

Check status

bash
# On each machine
./bin/devstation.sh status

# Via Telegram (from iPhone or any device)
/status

View logs

bash
# All services
./bin/devstation.sh logs

# Specific service
./bin/devstation.sh logs pipeline

Start / stop

bash
# On Mac Mini or Mac Studio
./bin/devstation.sh up
./bin/devstation.sh down
./bin/devstation.sh restart

Run the pipeline manually

bash
# On Mac Studio
docker exec devstation-studio-pipeline-1 python pipeline.py \
  --issue "AI-123" \
  --repo "user/repo" \
  --title "Fix login timeout" \
  --body "Description of the issue"

Use OpenCode on MacBook Air

bash
# Uses Studio's Ollama by default
opencode "fix the race condition in process.py"

# Override LLM source
opencode "add error handling" --model deepseek
opencode "refactor this module" --model claude

Remote control via iPhone

Open Telegram, message your bot:

CommandWhat happens
/statusBot replies with current pipeline state
/pausePipeline stops processing new tickets
/resumePipeline resumes
/approveCurrent task approved, PR is finalized
/rejectCurrent task rejected, healer retries
/llm cloudSwitch to cloud APIs
/llm studioSwitch to Studio's local Ollama

Backup & restore

bash
# On Mac Mini
./bin/devstation.sh backup

# Restore
./bin/devstation.sh restore backups/db_20260427_143022.sql

Update

bash
# Pull latest code + update Docker images
./bin/devstation.sh update

MIT Licensed | Built with AI Dev Station