Launching today
Every PaaS charges $5–7/month minimum to keep your container running. SnapDeploy charges $1 for 24 hours — no card, no auto-renewal. Push code from GitHub, we handle builds, SSL, and scaling on AWS. Free tier with 10 deploys/day. AI fixes broken Docker builds.
@yash_mahalle The "been burned by AI tools that touched things they shouldn't" framing
is exactly why SmartFix is scoped tight.
It only touches dependency manifests — requirements.txt, package.json, package-lock.json, pyproject.toml, Dockerfile. Never touches your application code (.py, .js, .go, etc.).
Catches roughly 7/10 dependency-class failures: missing package, wrong version pin, base image mismatch, missing apt dependency, Dockerfile syntax errors. Things where you'd otherwise have to read 200 lines of build log to spot the one bad line.
The instant we let it rewrite Python/JS/Go files, the trust contract breaks — so we deliberately don't. That's the bet.
Bookmarked. The "no Kubernetes, no YAML, no DevOps degree" framing is exactly what I need — last weekend I spent 4 hours trying to get a tiny FastAPI service deployed because I wanted to learn k8s "properly," and I just wanted the app online.
Two questions:
(1) what happens when the container OOMs — auto-restart or do I have to redeploy?
(2) any way to pipe container logs into something like Logtail or just use the dashboard?
@avishkar_jadhav223 That 4-hour k8s detour is exactly the pain this was built around — I went through the same loop with my own side projects before deciding to just
build the simpler thing.
To your questions:
(1) Auto-restart. ECS catches the OOM, replaces the task within 60-90s, no redeploy needed. The restart shows up in the Deploys tab with the OOM reason logged. You only have to redeploy if you actually need more memory or larger sized containers.
(2) Dashboard streams stdout/stderr live for now (right-click → copy raw
log). you can also check from the logs pafe with basic SnapDeploy filter options.
Native Logtail / Datadog / CloudWatch Logs forwarding is on the next-quarter roadmap.
For production observability today the workaround is wiring up your own log shipper inside the container.