# Jungle Grid

> Jungle Grid is a managed GPU execution platform for AI jobs.

Jungle Grid is a managed GPU execution platform for AI jobs. Submit inference, training, image, agent, and batch workloads without renting servers or choosing GPU providers.

## Facts
- Website: https://junglegrid.dev
- Pricing: free
- Categories: AI & Machine Learning
- LotsDirectory page: https://lots.directory/products/jungle-grid
- Listing state: Unclaimed. We listed this product from its public website, and its owner has not claimed it yet.

## Detailed info

## What is Jungle Grid?

Jungle Grid is a managed GPU execution platform that routes and runs agentic AI workloads through an MCP server, API, or CLI.

## What Jungle Grid does

The platform executes AI tasks including inference, batch processing, image generation, training, and containerized jobs without exposing raw server hardware. It performs intent-based capacity routing, tracks lifecycle stages from queued to completed, manages logs and output artifacts, and provides an integrated error diagnostic tool called Rex.

## Who Jungle Grid is for

Developers and AI teams running production workloads who want to submit jobs via MCP or API without managing GPU infrastructure.

## How Jungle Grid works

Applications or AI agents submit workloads via the Jungle Grid MCP server, REST API, or CLI. The platform finds suitable compute capacity, monitors execution progress, and returns logs, status updates, and output artifacts.

## Key features

- Intent-based routing allocating AI workloads across distributed capacity based on fit and reliability.
- Multi-interface access supporting submissions via MCP server, REST API, CLI, or web portal.
- Agent tool support connecting with coding assistants including Claude Code, Cursor, Codex, and Windsurf.
- Workload lifecycle management tracking jobs from queued to running, completed, or failed states.
- Rex diagnostic assistant that inspects workload failures and explains configuration issues.
- Artifact and log retrieval providing access to generated files and execution transcripts.

## Pricing

As listed on junglegrid.dev, checked 14 October 2026. Prices can change; see the site for current terms.

- Developer: $0 per month with 10 free execution jobs, free managed inference on supported models, and usage-based compute.
- Pro: $99 per month planned tier for production teams with Rex Pro diagnostics, higher concurrency, and extended limits.

## Facts from junglegrid.dev, checked 14 October 2026

- Managed execution platform for AI inference, batch processing, and training workloads.
- Provides workload execution via MCP server, API, and CLI.
- Features automated intent routing, lifecycle tracking, and Rex error diagnostics.
- Includes 10 free execution jobs on the $0 Developer plan.

Written from Jungle Grid's own website ([Homepage](https://junglegrid.dev/), [Pricing](https://junglegrid.dev/pricing), [Features](https://junglegrid.dev/how-jungle-grid-works), [Docs](https://junglegrid.dev/faq)). The maker can claim this listing to edit it.

## FAQ

### What types of AI workloads can run on Jungle Grid?
Jungle Grid supports inference, image generation, batch processing, model training, fine-tuning, and containerized AI jobs.

### How do AI agents interact with Jungle Grid?
AI agents connect through the Jungle Grid MCP server, allowing agents like Claude Code and Cursor to submit jobs, monitor status, and retrieve artifacts.

### Do developers need to select raw GPU hardware manually?
No. Jungle Grid is built around workload intent and automatically routes execution across suitable capacity without requiring manual GPU selection.

### Is there a free tier for testing workloads?
Yes. The Developer plan is $0 per month and includes 10 free execution jobs with no credit card required.

---
Source: https://lots.directory/products/jungle-grid · Machine-readable context by LotsDirectory.