Deploy and Host Aegra on Railway

Deploy Aegra 0.9.24 as an authenticated, self-hosted Agent Protocol backend with durable PostgreSQL checkpoints and Redis-backed job execution. The template includes a working echo graph, generated bearer authentication, automatic database migrations, and crash recovery without requiring an LLM key.

About Hosting Aegra

Aegra serves LangGraph applications through the Agent Protocol used by LangGraph clients. This Railway topology keeps PostgreSQL and Redis on the private network while exposing one managed-TLS API domain for threads, runs, streaming events, crons, and store operations.

Why Deploy Aegra on Railway

Railway supplies the public domain, private service discovery, generated secrets, and persistent volumes. Aegra starts its API and queued workers together, so a graph run can be accepted over HTTP, executed through Redis, checkpointed in PostgreSQL, and recovered after a process restart.

Common Use Cases

Host LangGraph agents behind the standard Agent Protocol.
Give internal applications a persistent threads-and-runs backend.
Test streaming, cron, store, and human-in-the-loop integrations on infrastructure you control.
Start from the included key-protected graph and replace it with an application-specific agent.

Dependencies for Aegra

Deployment Dependencies

Aegra CLI and API 0.9.24, installed from hash-locked official PyPI artifacts
pgvector on PostgreSQL 18 with a persistent volume
Redis 8 with password authentication, AOF persistence, and a persistent volume
No external model credential for the included echo graph