My AI Tech Stack: A Comprehensive Guide as we head into 2025
Ryan Boyle
As artificial intelligence continues to evolves at a rapid pace, I want to share my firsthand experience with the AI tools, services, platforms, and hardware that I use daily. This guide will help newcomers and AI enthusiasts understand which tools might best suit their needs—especially when working with a limited budget.
Note (March 9, 2025): A lot has changed since I first wrote this in early January. I've decided to keep this initial guide intact for historical reference. Moving forward, I'll publish monthly updates highlighting key changes, new tools, and updated insights—starting with January 2025 recap here.
Rating Criteria
I evaluate each tool based on:
Each tool receives a rating on a 5-star scale (⭐) and includes specific use cases for personal and/or professional applications.
Quick Reference Chart
Detailed Tool Reviews
1. Perplexity Pro (⭐⭐⭐⭐⭐)
Cost: $20/month Usage: Personal & Work Key Features:
If I could subscribe to only one AI tool, it would be Perplexity Pro. It excels at in-depth research, Q&A, finding code documentation, and more. It's completely replaced Google search for me, and I've changed my default search engine accordingly. The platform is intuitive yet powerful, especially with the Pro plan. I particularly value their Spaces feature for organizing research with files and links, and the ability to choose between different AI models (I primarily use Claude 3.5 Sonnet). I've never hit usage limits (at least with Sonnet 3.5) due to their very generous dynamic and rolling usage limits; it's effectively unlimited.
2. Claude (⭐⭐⭐⭐⭐)
Cost: $20/month Usage: Personal & Work Limitations: Low daily usage limits Strengths:
Claude emerged as the standout AI model of 2024. Since Opus 3.0's launch, it became my preferred model, and I immediately adopted the exceptional Sonnet 3.5 upon its release. Throughout the year, I used hundreds of millions of tokens or more across various tools - a testament to its remarkable capabilities. I'm eagerly anticipating their 2025 releases, especially Opus 3.5 and the 4.0 series of models.
Pro Tip: When working on longer projects in the Claude web or desktop app, occasionally ask Claude to provide a data dump of your work and then take that and start a new chat. This helps manage token usage as longer conversations consume more tokens and accelerate rate limit consumption. I often save Claude for my highest-leverage or most important daily tasks due to its restrictive usage limits. I would consider myself a power user, though, so you may have an easier time with the rate limits.
3. Windsurf (⭐⭐⭐⭐⭐)
Cost: $10/month base plan (early adopter pricing) Usage: Personal & Work Notes:
Windsurf delivers exceptional ROI through improved productivity and reduced mental overhead. It's become my preferred AI-native IDE, surpassing my previous heavy use of Cursor. The additional costs are justified by the productivity gains and reduced mental burden.
4. Cursor Pro (⭐⭐⭐⭐⭐)
Cost: $20/month Usage: Personal & Work Features:
I upgraded from the free plan as Pro became more cost-effective and unlocked additional features. For students or those just looking to cut costs, the BYO (bring your own) Google AI Studio - Gemini API key option continues to offer great value (free) with effectively unlimited usage. While I use it less than Windsurf now, it remains valuable as a complementary tool, especially when I hit roadblocks in Windsurf.
5. ChatGPT (⭐⭐⭐⭐)
Usage: Personal & Work Strengths:
While not my favorite overall, ChatGPT excels in specific use cases. Its data processing capabilities and voice features are particularly strong. The context window can be limiting compared to Claude, but I never hit usage limits on GPT-4o even with the most demanding workloads. Their o1 reasoning model series is innovative, impressive, and all signs point to AGI emerging from this line of work. OpenAI has done a lot of good work in 2024, but I expect what they release in 2025 will be earth shattering.
6. Raycast AI Pro (⭐⭐⭐⭐)
Usage: Personal & Work Key Features:
I use it tens, maybe even hundreds of times daily for quick writing improvements. The keyboard shortcuts make it extremely efficient, and the ability to make your own model selection (I chose Haiku 3.5 for speed, Sonnet 3.5 for quality) works perfectly for different use cases. It's an excellent backup when other tools' usage/rate limits are exhausted.
7. v0 (Vercel) (⭐⭐⭐⭐)
Usage: Work Plan: Pro Strengths:
As someone frequently working with frontend design and Next.js/React, v0 has been a game-changer. It excels at quickly bootstrapping new projects and features, making it invaluable for web development.
8. Cosmos (Humane AI Pin) (⭐⭐⭐⭐)
Usage: Personal Features:
My third favorite personal tool overall. The multi-step agentic capabilities and personalization are impressive, and the team's responsiveness to feedback has led to constant improvements.
9. GitHub Copilot (⭐⭐⭐⭐)
Cost: $10/month Usage: Work Features:
While the agentic workflows are just okay, the unlimited usage of powerful models at a reasonable price point makes it valuable, especially for coding assistance on smaller projects.
10. Lindy AI (⭐⭐⭐⭐⭐)
Cost: $49.99/month (Pro plan) Usage: Personal & Work Notes:
Though I'm new to it, Lindy AI has quickly proven its worth. The combination of time-saving features and productivity enhancements makes it a rising star in my toolkit.
11. Rewind AI & Limitless AI (⭐⭐⭐⭐½)
Cost: $20/month bundle (early adopter) Usage: Personal Features:
While Rewind's core product hasn't seen recent updates, it remains valuable to me for context retention and time tracking. The Limitless AI is an excellent platform-agnostic meeting recoder, tracscript generator and summarize, and I'm very much look forward to the pendant release this year.
12. Gemini (⭐⭐⭐½)
Usage: Personal & Work Strengths:
The models themselves (especially 1.5 flash and 2.0 exp) are impressive, but the product experience needs significant improvement. Current limitations with chat history retention and feature disparities between personal and enterprise versions make it difficult to recommend.
13. Notion AI (⭐⭐⭐½)
Status: Recently canceled Usage: Personal & Work Features:
While useful, Raycast AI now handles most of my Notion-related AI needs. It’s a good option for beginners, marketers, or heavy Notion users, but not essential for my workflow.
14. Replit Agent (⭐⭐⭐⭐)
Usage: Work Strengths:
Great for beginners and quick prototypes, but I prefer v0 for frontend design and Claude for other quick projects. The seamless/automatic hosting is a notable advantage for new developers.
15. Microsoft Copilot (⭐⭐⭐)
Usage: Personal & Work Plans: Tried both Personal and Enterprise Notes:
Despite Microsoft's access to strong models, the product itself doesn't justify its cost. I hope to see it improve in the future. I expect Gemini to surpass it in the near term, despite its own pitfalls.
16. Bolt (⭐⭐⭐⭐)
Usage: Personal & Work Features:
Shows potential but I need to spend more testing to have a fuller assessment. I currently prefer v0 for similar use cases, but I am planning to evaluate it more thoroughly in the future.
17. xAI - Grok (⭐⭐⭐)
Usage: Personal & Work Features:
The Twitter integration is interesting, but the model itself needs improvement, as it's not state of the art in any particular area. I'm excited to see what Grok 3.0 brings to the table.
18. Zed (⭐⭐⭐⭐)
Usage: Personal & Work Features:
Great potential with impressive technical foundations, but lacks agentic workflows. The context window length is outstanding, and being open source is a plus. It just hasn’t ‘clicked’ for me yet. Looking forward to testing it again soon on more serious projects to see if my mind changes on this.
19. Apple Intelligence (⭐⭐)
Usage: Personal Features:
Basic but reliable for simple tasks. My usage of it is currently limited to notification summaries. It desperately needs feature expansion to become more useful in my workflow, but the privacy-first approach is commendable.
20. Rabbit R1 (⭐⭐)
Usage: Personal Hardware Cost: $199 Status: Inactive (in drawer after 2-week trial) Features:
Despite the bad press & reviews it received on launch, it has improved a lot via software. It would be interesting to see tech & product reviewers give it another chance in 2025. While the concept is innovative and the initial setup was straightforward, the R1 hasn't found a permanent place in my daily workflow. After giving it a ~two-week trial, it's now sitting unused. I find myself preferring the Humane Ai Pin (Cosmos) for portable AI assistance, particularly due to their stronger data security practices and more polished user experience. However, I remain optimistic about Rabbit's potential and look forward to seeing how the platform evolves. The Large Action Model (LAM) concept is interesting, and with continued development, it could become a more competitive option in the portable AI assistant space. Although, they may need to build an app (probably with a subscription) to stay relevant long term, I don’t think that’s a bad thing, but I’m not sure if that’s the route they want to go.
Budget Recommendations
Essential Tier ($35/month)
This combination provides a robust foundation, giving you powerful research capabilities and an agentic (they call it Cascade) AI-powered development environment. These two tools alone can cover most essential needs for both personal and professional use.
Enhanced Tier (Additional $20-60/month)
Note: Many users find Claude in Perplexity Pro sufficient for their needs, so consider your use cases carefully. For instance, if you rarely use voice features, you might skip ChatGPT and invest in Cursor Pro instead.
Professional Tier (Additional $90-100/month)
This tier significantly enhances your professional workflow with specialized tools for development, productivity, and context awareness.
AI Early Adopter Tier (Additional Hardware Costs)
This tier is for those who want to be at the cutting edge of AI hardware and software integration. The Humane AI Pin adds powerful portable AI capabilities to your existing stack.
Pro Tip: Start with the Essential tier and gradually add or swap tools based on your specific needs and workflow bottlenecks. Many users find they can accomplish most tasks with just the Essential tier tools, especially when maximizing their capabilities. If used correctly, all AI tools should more than pay for themselves, if they aren’t, consider if the tool is right for your workflow or if there are ways to enhance and expand your usage of them.
Usage Strategies
For Maximum Productivity:
For Token Management:
Looking Ahead
The AI landscape continues to evolve rapidly. I'm particularly watching:
A closing note
Your ideal AI stack will depend heavily on your specific needs, budget, and workflow. I hope sharing my experiences helps you make more informed decisions about which tools to invest in. Remember that while having access to multiple AI tools can be powerful, starting with one or two essential services and gradually expanding based on concrete needs often works best.
I'll continue to post updates as tools evolve and new options emerge. Feel free to reach out with questions or share your own experiences with these tools - I’m always excited to talk about AI.
Consider subscribing to my email newsletter to stay updated on new posts. I promise to keep emails minimal, only notifying you about new content and significant updates to existing articles like this one.