Midrender

Active

Turn product features into launch videos

Summer 2023Founded 20231 peopleSan Francisco, CA, USA

midrender.com ↗LinkedIn ↗X ↗GitHub ↗See on the Idea Map B2B momentum

About

Midrender is a visual editor for creating motion graphics, powered by AI. Point the agent at your codebase, it reads your code, understands the feature, and creates an animated walkthrough you can ship. A full timeline and design editor lets you fine-tune every detail.

From their website

as of Jun 7, 2026midrender.com ↗

SaaSSubscription · Free tier available with 3 projects and 100/week MCP tool calls; Pro at $20/month with 1M/week MCP tool calls; Business at $100/month with unlimited collaborators; Enterprise pricing based on company size and volume usage.

Midrender enables AI-assisted motion graphics where AI agents and humans collaborate to create animations. It offers a visual editor, integration via CLI/MCP, and tools to keep designs on-brand from a shared source of truth.

Users describe target motions in plain language, an agent generates assets and animation, and everything remains fully editable by humans. The product includes a visual editor for fine-tuning timelines, properties, and keyframes, supports integration through a CLI or MCP to fit into existing workflows, and can enforce design-system constraints by reading components, typography, and style guidelines. It also offers embedding capabilities for web use and separate pricing tiers including Free, Pro, Business, and Enterprise with usage allowances and collaboration features.

Who it’s for: Teams and agents who need rapid motion graphics production while maintaining manual editability and on-brand output; organizations seeking to embed animation creation into their workflow through CLI/MCP and design-system integration.

Features

plain-language motion description to generated animation
full visual editor for timeline, properties, and keyframes
human-agent collaboration on the same source of truth
design-system matching via components, typography, and style guidelines
CLI and MCP integrations to embed into existing stacks
web embedding capability (coming soon)

Pricing page with multiple tiers and free trial; mentions of beta/product updates; mentions of embedding features and CLI/MCP integrations indicating active development and traction.

Founders · 1

Konstantin HohrFounder

Konstantin is the founder of Revideo. Previously, he studied computer science at RWTH Aachen while simultaneously working as a software engineer at Jodel, where he built content moderation and abuse detection systems. At DynamoAI (W22), he spearheaded development on vertical federated learning based products.

LinkedIn ↗X ↗

Launch

Launched on Y Combinator · Aug 2023

View launch post ↗

Haven makes it easy to deploy LLMs on your own infrastructure

Haven provides a Kubernetes-based service to deploy and manage LLMs on customers’ own AWS or GCP infrastructure, offering serverless-like deployment experience while keeping data on-premises. Users upload a service account key, configure model/GPUs/scaling, and run inference with fast performance and features like input batching.

From the original launch (Aug 2023) — may be outdated.