Nicolas Coutureau
Nicolas Coutureau is an independent developer who builds and runs KinoPipe, an FFmpeg-as-a-service platform for AI agents. KinoPipe gives agents typed, validated video-editing operations through MCP or REST instead of exposing a shell, then returns a finished downloadable file.
At a Glance
- AI-agent builders and developers
- Backend/API teams needing hosted FFmpeg without operating video infrastructure
- Creators and social-media teams producing short-form video
- Automation users and n8n workflow builders
- +1 more
AI Tools by Nicolas Coutureau
(1)
KinoPipe
FFmpeg API for AI Agents
Discussions
No discussions yet
Be the first to start a discussion about Nicolas Coutureau
Latest News
KinoPipe MCP repository added a Claude skill with read-before-edit guidance and analysis tools.
KinoPipe published its FFmpeg 9 serverless CPU-versus-GPU benchmark report, including 325 RunPod runs and 95 Cloud Run runs.
KinoPipe MCP README updated with 43 typed tools, benchmark results and integrations.
Public KinoPipe MCP listing and integration repository published.
Products & Services
Hosted FFmpeg 9 video-processing service for AI agents, developers and automations. It exposes 43 typed tools and custom multi-operation jobs through REST and a hosted MCP server; jobs are asynchronous and produce temporary signed download URLs.
Public remote MCP integration for KinoPipe at https://kinopipe.com/mcp, with OAuth 2.1 or bearer-key authentication, typed tool discovery, job polling and downloadable outputs. The public GitHub repository contains the server listing, integration documentation and Claude skill; the processing source is private.
An Apify integration authored by Nicolas Coutureau that runs 22 typed KinoPipe editing tools, validates inputs, submits and polls asynchronous jobs, and returns temporary signed output URLs. It supports usage-based video-credit billing at $0.01 per credit.
Market Position
KinoPipe positions itself as a typed, safety-focused alternative to shell wrappers and general video APIs: it validates bounded operations for AI callers, supports MCP, and compiles compound edits into a single FFmpeg pass. Its own comparison pages distinguish it from Rendi (raw FFmpeg commands), Shotstack (timeline composition) and Creatomate (template generation), positioning KinoPipe around finishing and editing existing media for agents.
Leadership
Founders
Nicolas Coutureau
Independent developer based in France and the sole builder/operator identified by KinoPipe. His public profile describes him as a software engineer building Revid.ai and previously at Jitter Video; Revid identifies him as CTO and co-founder responsible for its AI video pipeline.
Executive Team
Nicolas Coutureau
Founder and operator, KinoPipe; independent developer
Software engineer and independent developer in France. He previously worked with Jitter Video and is CTO and co-founder of Revid.ai, where Revid says he built the technical foundation for its AI video pipeline.
Founding Story
KinoPipe began from Coutureau's observation that AI agents are good at deciding what video edit is needed but unsafe at running FFmpeg directly. He started KinoPipe to provide a bounded, typed contract with validation, asynchronous jobs and an MCP server, rather than command-line access; the site describes it as new and independent.
Business Model
Revenue Model
Credit-based SaaS and usage pricing. Customers pay for successful processing based on the greater of rounded-up worker seconds or rounded-up 6 MiB output units; subscriptions provide monthly credits, and a one-time credit pack is also available. Failed jobs are refunded automatically.
Pricing Tiers
100 credits once, approximately 10 edits; complete video API; no credit card required.
1,000 credits, approximately 100 edits; credits never expire while the account is open.
1,500 credits per month, approximately 150 edits.
5,000 credits per month, approximately 500 edits; marked most popular.
25,000 credits per month, approximately 2,500 edits; priority support.
Target Markets
- AI-agent builders and developers
- Backend/API teams needing hosted FFmpeg without operating video infrastructure
- Creators and social-media teams producing short-form video
- Automation users and n8n workflow builders
- Prototype, side-project and production agent workflows
- AI agents performing video edits through natural-language requests
- REST API video processing in backend applications
- Claude, ChatGPT, Codex, Cursor, GitHub Copilot and Windsurf agent workflows
- n8n automations and other developer workflows
- Social-video production for Instagram Reels, TikTok and YouTube Shorts
- Pre-edit analysis of long recordings to find scenes, silence, highlights and captions