Living resource · Updated 2026-07-04

My AI Stack

The living list of AI tools I personally use, pay for, canceled, switched from, and recommend.

Personally testedEditor's pickEditor favorite

How the stack works together

One general assistant, one research layer, one creation layer, then automation only after the workflow is proven.

1

Think

2

Create

3

Systemize

Paid tools

5

Subscriptions stay only if they earn a clear role in the weekly workflow.

Get stack updates

Stack by workflow area

Writing · Daily

Claude

It is the tool I trust most for thoughtful drafts, editing, positioning, and long-form reasoning.

4.5
What I replaced
Most standalone AI writing apps
What I love
It handles nuance better than generic writing templates and rarely needs a heavy rewrite.
Downsides
It does not generate native images and can hit usage limits on heavy writing days.

Research · Daily

Perplexity AI

Citations make it faster to move from a question to a source-backed working brief.

4.4
What I replaced
Manual Google tab sprawl
What I love
It keeps sources visible, which makes verification easier.
Downsides
It is a starting point, not a substitute for reading the underlying sources.

Coding · Daily

Cursor

Codebase-aware edits are more useful than simple autocomplete when building real features.

4.5
What I replaced
Generic coding autocomplete
What I love
It can reason across multiple files and turn rough implementation notes into working changes.
Downsides
It can be too aggressive, so every diff still needs review.

Marketing · Weekly

Canva AI

It helps create practical campaign assets without turning every visual request into a design project.

4.3
What I replaced
Ad hoc one-off design requests
What I love
Templates keep repeatable assets fast and consistent.
Downsides
Advanced creative direction still needs a designer's eye.

Video · Project-based

Descript

Transcript-based editing makes spoken content much easier to clean up and repurpose.

4.5
What I replaced
Manual audio/video cleanup
What I love
Editing a recording feels closer to editing a document.
Downsides
Complex productions still need dedicated editing workflows.

Images · Weekly

Midjourney

It is still my favorite when visual quality and style matter more than perfect interface polish.

4.7
What I replaced
Stock image searches for concept visuals
What I love
The output has a stronger creative point of view than most image tools.
Downsides
The workflow is less approachable than Canva or Adobe tools.

Automation · Project-based

Zapier AI

It is useful once a process is proven and repeated often enough to deserve automation.

4.3
What I replaced
Manual handoffs between apps
What I love
It connects everyday business tools without custom engineering.
Downsides
Automating a messy process usually makes the mess faster.

Meetings · Weekly

Fathom

It captures decisions and follow-ups so meetings produce usable records.

4.5
What I replaced
Handwritten meeting summaries
What I love
It reduces the admin burden after calls.
Downsides
Sensitive meetings still need consent and human judgment.

Productivity · Weekly

Notion AI

It keeps notes, decisions, project context, and lightweight AI help in one workspace.

4.2
What I replaced
Disconnected notes and scattered prompts
What I love
It is strongest when the workspace already contains useful context.
Downsides
It is not the best standalone AI assistant if your knowledge base is thin.

Paid Subscriptions

Daily Workflow

Claude

ChatGPT

Perplexity AI

Cursor

Weekly Workflow

Canva AI

Descript

Midjourney

Notion AI

What I Canceled

Extra single-purpose writing apps

Claude and ChatGPT covered most writing workflows with less tool sprawl.

Overlapping meeting bots

One reliable meeting assistant is enough for most teams.

What I Switched To

Generic coding autocompleteCursor

Codebase-aware edits matter more than single-line suggestions for larger changes.

Manual research tabsPerplexity AI

Cited synthesis is faster for early research and comparison briefs.