Skip to main content
AIDiveForge AIDiveForge

Cognikernel vs Ocdify

Nothing in common beyond both running locally. CogniKernel is a developer tool that gives Claude Code and Codex a memory of past decisions on a project. OCDify is a Mac menu-bar app that sorts your Downloads and Desktop into folders using an on-device model. The few people who reach this page searched for CogniKernel, so most of this is about that, with a note that its author has paused updates.

CogniKernel, for coding agents

CogniKernel records the decisions, constraints and abandoned approaches from your coding sessions and feeds them back later, so the agent doesn't reopen settled questions. In Claude Code it uses four hooks: it injects memory at session start, surfaces relevant notes as you prompt, warns before a tool call breaks a past decision, and extracts new decisions when a session stops. Codex has no hooks, so it reads memory through CogniKernel's MCP server and you sync with cognikernel codex-sync. Both tools share one store per project.

The store is a local SQLite event log with full-text search, and optional embeddings if you install the [embedding] extra. It's Apache 2.0 and installs with pip, pipx or uv. The README now says updates are paused until further notice, with v0.1.3 (September 2026) staying current.

OCDify, for a messy Mac

OCDify watches folders you pick, reads each file with a local model and moves it into categories you name. There's an optional Smart Rename for files called things like scan001. The App Store listing asks for an M1 or later Mac, notes a model of about 1 GB downloaded on first launch, and gives a price of $9.99, paid once. The vendor's site also sells a direct license, and offers a 30-day trial with no card.

Cognikernel: Pick CogniKernel if you run Claude Code or Codex on long projects and keep re-explaining the same decisions, and you're fine using a project whose updates are paused.

Ocdify: Pick OCDify if you have an Apple Silicon Mac and want Downloads and Desktop sorted automatically without sending files to a server.

Facts checked Oct 18, 2026 against 4 sources

Which one should you try?

Cognikernel

Free

Best for Users running Claude Code or Codex on multi-session projects

Ocdify

Freemium

Best for Mac users who prioritize data privacy and offline processing

Pricing, capabilities and platforms side by side. Prices are as of the "Pricing checked" date in the table.

Cognikernel

Cognikernel

The tool hooks into Claude Code and Codex session surfaces, extracts decisions, constraints, and discarded approaches, and writes them into an event-sourced log keyed on the project path — so the next session picks up where the last one stopped. Because the store is path-keyed and local, memory made in Claude Code is readable by Codex on the same project without any sync step. There is no vector database, no embeddings infrastructure, no API call — just typed, auditable memo records on disk. The ceiling appears when your context needs go beyond structured decisions: narrative code understanding, semantic search across past sessions, or anything requiring retrieval ranked by similarity will not work here.

Ocdify

Ocdify

The core loop is simple: point OCDify at a directory, define your category names, and let it classify and move files in the background while you work. Because the AI runs entirely on your Mac, nothing leaves your machine — a real constraint lifted for anyone handling client documents, medical records, or anything you would not want passing through a third-party API. The menubar footprint keeps it out of your way. The ceiling appears when your organization logic gets genuinely complex: OCDify applies categories you name, but it does not negotiate ambiguity, rename files according to project conventions, or move items across drives. Teams that need rule-based routing with exceptions built in will hit that wall fast.

AttributeCognikernelOcdify
PricingFreeFreemium
Price—$9.99 one-time
Pricing checkedNot checked yetOct 19, 2026
Free trialNot stated30 days
Open sourceYesNo
Has APINoNo
Self-hosted optionYesYes
PlatformsPython package (pip/pipx/uv); Claude Code hooks and MCP server for CodexmacOS (Apple silicon, M1 or later)
Pros
  • Event-sourced, typed decision log so every constraint the agent is told about is inspectable and version-controllable — meaning you can audit exactly what context shaped a session instead of trusting a black-box embedding store.
  • Project-path-keyed storage, so memory written during a Claude Code session is automatically available in a Codex session on the same project — eliminating the copy-paste handoff developers otherwise do manually between tools.
  • Fully local, no-API, no-server architecture, which means there is no per-token cost for memory operations and no external dependency that breaks when an API rate-limits you mid-session.
  • Fail-open design described by the vendor, so a missing or corrupt memory store does not block the coding session — the agent continues without context rather than erroring out.
  • Apache-2.0 license with self-hosted-only deployment, so the memory store never leaves your machine and is not subject to a SaaS vendor's data retention or privacy policy.
  • Fully offline AI classification using a local LLM, which means files containing confidential or personally identifiable information never leave your Mac — eliminating the compliance risk that comes with cloud-connected file tools.
  • Continuous directory watching rather than on-demand sorting, so folders stay organized passively without requiring you to remember to run anything.
  • Custom category definitions you control, which means the AI sorts into folder names that match your actual workflow rather than a vendor-imposed taxonomy you have to retrofit.
  • One-time purchase with lifetime updates (vendor-stated), so there is no recurring billing to cancel, audit, or justify in a budget review.
  • Menubar-only interface, which means it runs without occupying dock space or requiring an open window — a real difference when screen real estate is already allocated.
Cons
  • The tool captures structured decisions and constraints, not semantic understanding of code — so when you need to ask 'find past sessions where we discussed authentication' and rank results by relevance, there is no retrieval mechanism for that. Teams with those needs add a vector store alongside CogniKernel, at which point they are maintaining two separate memory systems.
  • Hook integration is limited to Claude Code and Codex surface exposure — any coding assistant that does not expose a hook interface gets no memory injection, which forces teams running mixed toolchains to switch to a competitor with broader IDE or assistant integrations.
  • There is no API surface, so automated pipelines or CI steps that need to read or write to the memory store must interact with the file format directly. Teams building agent orchestration around this will be writing their own integration glue rather than calling a documented endpoint.
  • OCDify categorizes and moves files, but does not rename them. If your workflow depends on enforced file naming conventions — project codes, date prefixes, client identifiers — you will still need a separate tool or manual cleanup after every sort.
  • There is no API and no scripting surface exposed, based on the page. When your organization logic grows beyond 'assign to one of these categories,' there is no extension point — you cannot add conditional rules, override classifications for specific file patterns, or pipe output into another system. Teams that reach this ceiling switch to Hazel or a custom AppleScript setup.
  • The tool is Mac-only with no self-hosted server option or cross-platform support indicated. Any team member on Windows or Linux is excluded from the workflow entirely, making this a poor fit for mixed-OS environments where shared folder organization is the actual problem to solve.

Made up your mind?

Cognikernel

Free

Best for Users running Claude Code or Codex on multi-session projects

Ocdify

Freemium

Best for Mac users who prioritize data privacy and offline processing

Specs come from each tool's own site and docs (how we test). AIDiveForge is editorially independent.