Developer docs

Getting started

Every project lives in the dashboard at app.kre8ivhosting.io. Connect a Git repository or upload a container image, choose a plan, and we build and deploy it. Environment variables, domains and certificates are managed from the same place.

Deploying an app

Push to the tracked branch and the platform builds a new image, deploys it next to the current version and switches traffic once it is healthy. Every build is kept, so you can roll back to any earlier version.

git push origin main
# build → health check → switch traffic
# roll back from the dashboard: Deployments → … → Roll back

Model gateway

One endpoint and key for OpenAI, Anthropic and Amazon Bedrock models. Set monthly limits per key, see every request in the logs, and switch providers without changing your code. The gateway accepts OpenAI-style requests.

curl https://gateway.kre8ivhosting.io/v1/chat/completions \
  -H "Authorization: Bearer $K8_GATEWAY_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-sonnet","messages":[{"role":"user","content":"Hello"}]}'

Vector databases

Every hosting plan includes managed PostgreSQL with the pgvector extension enabled. Store embeddings next to your application data and query them with SQL.

CREATE TABLE docs (id bigserial PRIMARY KEY, body text, embedding vector(1536));
SELECT id, body FROM docs ORDER BY embedding <=> $1 LIMIT 5;

GPU instances

GPU plans run on dedicated NVIDIA instances. Attach one to a service for inference, or run scheduled jobs for batch work and fine-tuning. GPU hours are metered and shown on your usage page in real time.

Assistants and RAG

Managed assistants are configured for you. You choose the sources (website, files, drives), the tone and the handoff rules. We handle chunking, embeddings, retrieval and quality checks, and review unanswered questions with you monthly.

Security and data

Data is encrypted in transit and at rest. Model providers are used under business terms that don't train on your data by default. Access to your environment is limited to the engineers on your account and is logged.

Billing and usage

Plans are billed monthly through the Kre8ivHosting client area. Model usage and GPU hours are billed separately at the end of each month, with alerts at the limits you set.