ClineFlow · Open Knowledge Format

ClineFlowPersistent Context.
Open Knowledge.

Infinite AI Memory across Chats, Agents and Teams.

AI coding memory, now native OKF.

01 / Ground Define the work.
SpecificationsGoalsDecisions
02 / Verify Make success observable. Plans → Tests → Verification
03 / Carry forward Knowledge that stays. “please commit” A journal. A verified handoff.

Try the Agentic Installer

Run this command from your project folder using your preferred AI agent.

Please install ClineFlow by following the instructions provided at https://github.com/hassanvfx/clineflow

ClineFlow compatibility with Claude, OpenAI Codex, Cursor, Cline, and GitHub Copilot
ClineFlow is supported by all major Agents

The ClineFlow loop

It’s easy as 1–2–3.

Keep the conversation natural. ClineFlow handles the durable project context behind the scenes.

  1. 01

    Install once.

    Paste the Agentic Installer prompt above. It installs ClineFlow and configures the project for you.

  2. 02

    Describe the outcome.

    Tell the agent what you want to build. ClineFlow automatically maintains the specs, goals, verification, decisions, and journal.

  3. 03

    Say “please commit.”

    After you verify the result, close the loop. ClineFlow journals the verified state so the next chat, agent, or teammate can continue.

Learn More

Install from GitHub

ClineFlow: Agentic AI Knowledge Base

100% Automatic Context Journaling

* Google’s Open Knowledge Format Compatible

The Problem

AI Context Is an Open Problem in Agentic Development

Models are becoming more capable, but project context is still fragile. When a chat closes or an agent changes, plans, constraints, decisions, and verification history can disappear with it. ClineFlow makes that working knowledge durable, portable, and automatic.

Examples of the wider conversation around durable agent context and memory.

ClineFlow Masterclass

Watch the Full 30-Minute Masterclass

Why this matters

Your agent does not need another database. It needs project knowledge.

AI context and memory is emerging as one of the central problems in agentic development. Vendor-specific memory can be powerful, but it can also tie a project’s history to a single ecosystem. ClineFlow takes a different path: it gives agents a local knowledge base beside the project.

Specifications, goals, plans, tests, decisions, and verification history are more than application data—they are the working knowledge that guides the work. Stored with the project, that knowledge is portable, versionable, collaborative, and available across chats, agents, and handoffs through familiar Git workflows.

This masterclass shows how that principle becomes an automatic development loop.

ClineFlow Workflow

From a goal to a durable handoff.

Follow the complete automatic development loop—from grounding the work and verifying it to preserving the next agent-ready step.

Full Workflow
Cover of Infinite AI Context: ClineFlow and Google’s Open Knowledge Format by Hassan Uriostegui

New Book · 2026

Infinite AI Context

ClineFlow and Google’s Open Knowledge Format

Build durable AI memory for developers, lawyers, and creatives.

A practical guide to persistent AI context, ClineFlow, and Google’s Open Knowledge Format—designed for people building durable memory across development, legal, and creative work.

Install from GitHub

Hassan Uriostegui

About the Author

Hassan Uriostegui

Hassan Uriostegui is an AI-native principal engineer, founder, and author building agentic systems, consumer products, and mobile platforms. He created ClineFlow to give AI agents durable, auditable project memory across chats, tools, and collaborators.

His work spans context engineering, iOS, real-time systems, and generative workflows—from 0-to-1 products to platforms used by millions.

Visit Hassan’s portfolio