Hot Dev – Backend Workflows and AI Agents in One Open-Source Platform
Hot Dev is an open-source platform for building backend workflows—events, schedules, AI agents, and MCP tools—with a custom language, local dev runtime, and one-command deploys.
TL;DR
TL;DR: Hot Dev is an open-source platform for backend workflows—events, schedules, AI agents, and MCP tools—built around a custom language called Hot, with a local dev runtime and one-command production deploys.
What Is Hot Dev?
Hot Dev is an open-source platform for building backend workflows and AI agents. The project describes itself as:
“Open source platform for backend workflows: events, schedules, AI agents, MCP tools, long-running tasks, and service orchestration. It includes execution tracing, a local dev runtime, and single-command deploys.”
The core is a language called Hot, shipping as a compiler, VM, and standard library. On top sits the full platform: a CLI, REST API, web app, scheduler, event worker, task worker, and LSP server. The public Hot packages live under hot/pkg, including hot-std and provider or tool integrations.
License: Apache-2.0 (verified from the LICENSE file in the repository).
Setup Workflow
Prerequisites
- macOS, Linux, or WSL2
curlfor the installer- Or
cargoif building from source
Step 1: Install the CLI
The quickest path is the official installer script:
curl -fsSL https://get.hot.dev/install.sh | sh
This installs the hot binary globally.
Step 2: Build from Source (optional)
If you prefer building from the repo:
git clone https://github.com/hot-dev/hot.git
cd hot
cargo build --release --bin hot
# Binary lands at target/release/hot
Step 3: Create Your First Workflow
A Hot workflow file (.hot) wires webhooks, events, schedules, and MCP tools through a shared meta mechanism:
::myapp ns
::http ::hot::http
::uri ::hot::uri
// Receive a webhook, then fan out through an event.
on-signup
meta {
webhook: {service: "leads", path: "/signup"},
on-event: "lead:new",
}
fn (request) {
send("lead:new", request.body)
{ok: true}
}
// React to the event: score the lead and route it.
qualify-lead
meta {on-event: "lead:new"}
fn (event) {
score score-lead(event.data)
if(gte(score, 0.7),
send("lead:qualified", event.data),
send("lead:nurture", event.data))
}
// Run on a schedule.
weekly-summary
meta {schedule: "every monday at 9am"}
fn (event) {
post-pipeline-summary()
}
// Expose a function as an MCP tool.
get-forecast
meta {
mcp: {
service: "weather",
description: "Get the weather forecast for a location",
},
}
fn (location: Str): Vec {
loc ::uri/encode(location)
response ::http/get(`https://wttr.in/${loc}?format=j1`)
response.body.weather
}
Step 4: Run Locally
hot run myapp.hot
The local runtime starts the event worker, scheduler, and HTTP listener in one process. Changes to .hot files reload on save.
Step 5: Deploy to Production
hot deploy
This connects to Hot Dev Cloud and pushes your workflow. The platform handles event routing, scheduled runs, and MCP tool hosting.
Deeper Analysis
The Hot Language
Hot is a domain-specific language designed for backend orchestration. It uses a functional style with explicit metadata blocks (meta {}) for declaring triggers—webhooks, events, schedules, or MCP interfaces. This separates the “what triggers this” from “what this does”, making workflows easier to reason about.
MCP Tool Integration
The MCP (Model Context Protocol) integration lets you expose any Hot function as a tool an AI agent can call. The example above shows a get-forecast function exposed as an MCP tool under the weather service. This is a practical pattern for giving AI agents real-time data access without building a custom API layer.
Event-Driven Architecture
The send / on-event pattern lets workflows fan out from a single trigger. A webhook fires once; the event bus dispatches to any number of handlers independently. This is the same model as Cloudflare Workers’ queue + consumer or AWS EventBridge, but self-hosted and language-native.
Execution Tracing
The platform includes execution tracing out of the box. Each run records inputs, outputs, and timing. This is valuable for debugging event-driven workflows that span multiple handlers.
Practical Evaluation Checklist
- [ ] Install the CLI (
curl -fsSL https://get.hot.dev/install.sh | sh) - [ ] Clone the repo and build from source
- [ ] Run a webhook-triggered workflow locally
- [ ] Wire two handlers to the same event
- [ ] Deploy with
hot deploy - [ ] Verify tracing output in the web app
Security Notes
- Hot is open source under Apache-2.0. Audit the source at github.com/hot-dev/hot before running in production.
- The local runtime executes Hot code in the VM. Network access, file access, and external tool calls are controlled by the standard library (
::hot::http,::hot::uri, etc.). - For self-hosted deployments, the event worker and scheduler run as long-lived processes—apply standard OS-level hardening.
FAQ
Q: Is Hot Dev tied to Hot Dev Cloud, or can I self-host everything? A: The core Hot language, compiler, VM, CLI, API, scheduler, event worker, task worker, and LSP are all open source and self-hostable. Hot Dev Cloud is the hosted deployment option.
Q: What languages can I use for custom tools and integrations?
A: There are official Hot packages for JavaScript/TypeScript (hot-js), Python (hot-python), and Go (hot-go), plus demo repositories for each.
Q: How does Hot compare to Trigger.dev or Inngest? A: Hot is a custom language (not a YAML/JSON DSL), supports MCP natively, and is fully self-hostable. Trigger.dev and Inngest are hosted-first. If you want full control over your event infrastructure without learning a new general-purpose language, Hot is a closer match.
Q: Does it work on Windows? A: The CLI targets macOS and Linux directly, with WSL2 as the recommended Windows path.
Source and Accuracy Notes
- Project page: hot.dev
- Source repository: github.com/hot-dev/hot
- License: Apache-2.0 (verified from LICENSE file in the repository)
- Install script: get.hot.dev/install.sh
- HN launch thread: news.ycombinator.com/item?id=46797907
- Source last checked: 2026-08-07
Related Posts
dev-tools
Automotive Skills Suite for AI Engineering
Evaluate Automotive Skills Suite for APQP, ASPICE, HARA, safety-plan, and DIA workflows with setup notes, governance risks, and SME review guidance.
5/28/2026
dev-tools
awesome-agentic-ai-zh Roadmap Guide
Explore awesome-agentic-ai-zh as a Chinese agentic AI learning roadmap, with setup notes, track selection, study workflow, and evaluation guidance.
5/28/2026
dev-tools
Baguette iOS Simulator Automation Guide
Set up Baguette for iOS Simulator automation, web dashboards, device farms, gesture input, streaming, and camera testing with Xcode caveats.
5/28/2026