Conjure
Prompt-to-Infrastructure

From chat
to running
infrastructure.

Chat freely about your infrastructure. Conjure builds the architecture diagram, writes Terraform HCL, and provisions it — all from one conversation.

Free to start — no credit card required.

conjure.app/session
AWS Production Stack
AWSterraform
Add a Redis cache between the API servers and the database
Added Redis cache (cache.t3.micro) connected to your API servers on port 6379. Config updated.
↗ diagram updated
Load BalancerAPI ServerAPI ServerRedis CachePostgreSQL

01 / How it works

From conversation to cloud

01

Chat

Describe your infrastructure in plain English. Or link a GitHub repo with existing Terraform files to reverse-engineer into a diagram.

02

Diagram

A live Mermaid architecture diagram updates with every message. Click any node to edit its configuration in a properties drawer.

03

Generate

Click "Generate Code" to produce Terraform HCL — main.tf, variables.tf, outputs.tf — ready to download as a .zip or push to GitHub.

04

Deploy

Run terraform plan and apply directly from the browser. Target AWS or GCP using saved credential profiles.

02 / What you get

Everything in one conversation

Conversational design

Iterate through chat, not forms. The AI classifies every message and decides what to update.

Live architecture diagram

Mermaid-based, renders in real time as you chat. Click any node to view and edit its config inline.

Dual source of truth

Mermaid captures topology. YAML captures everything else — instance sizes, ports, networking. Always in sync.

Multi-provider

AWS and GCP supported. EC2, RDS, ElastiCache, Cloud Run, Cloud SQL, and more.

Terraform HCL output

Syntax-checked before display. Download a ready-to-run .zip or push the generated code to your GitHub repo.

Import existing repos

Link a GitHub repo with .tf files. Conjure reverse-engineers them into a diagram and config — no starting from scratch.

03 / Models

Start immediately. Upgrade when you want.

Conjure ships with built-in LLM models via OpenRouter — no API key needed. Add your own Anthropic key in Settings to unlock Claude.

Built-in

No key needed

Powered by OpenRouter

  • Nemotron Super 120B
  • GPT OSS 120B
  • Unlimited sessions
  • AWS & GCP support
  • Full diagram, code & deploy

Claude

Bring your key

Your Anthropic API key

  • Everything in Built-in
  • Claude Haiku 4.5
  • Claude Sonnet 4.6
  • Claude Opus 4.6
  • Higher quality outputs

Ready?

Start building your infrastructure

Free models, no credit card, no lock-in.
Just describe your cloud.