[ DEPLOY // TARGET INDEX ]
Mister Morph Deployment Index
Treat this page as a deployment router: decide by operational priority, then jump into the target playbook.
30-second target picker
| If your top priority is… | Pick | Why |
|---|
| Shipping a Telegram bot quickly on AWS | AWS Lightsail Containers | Fastest path to first deploy with a container-first workflow |
Exposing serve mode over public HTTP with edge ingress | Cloudflare Worker + Container | Edge entrypoint + container runtime split is built for this |
| Owning every layer (host, process, filesystem, hardening) | systemd on VM/server | Maximum control, lowest platform lock-in |
Deployment tracks at a glance
AWS Lightsail Containers
- Best when: you want a practical production-ish deploy soon, especially for Telegram long polling.
- Operational profile: AWS-native IAM/S3 workflow, low ceremony rollout.
- Watch out for: AWS coupling and less flexibility than a self-managed VM.
Cloudflare Worker + Container
- Best when: you need edge ingress in front of a containerized
serve endpoint. - Operational profile: split model (Worker + container + Wrangler), global edge entry.
- Watch out for: more moving parts and tighter Cloudflare platform coupling.
systemd on VM/server
- Best when: you optimize for control, auditability, and host-level hardening.
- Operational profile: direct Linux service management with full tuning freedom.
- Watch out for: you own patching, uptime, ingress, TLS, and scale mechanics.
Preflight checklist (before any target)
- Lock your provider/model/api key strategy (
MISTER_MORPH_* or config file). - Decide where secrets live and how rotation will be handled.
- Confirm persistence for logs, dumps, and any cached files you need.
- Define restart and failure behavior (crash, network outage, API timeout).
- Set minimum observability: startup logs, error logs, and health probes.
- Decide your allowlist boundaries for Telegram/Slack and inbound HTTP.
Entry docs
Production note
After choosing one track, execute only that track’s README end-to-end first. Avoid mixing deployment patterns in the initial rollout.