I am back at square one and I am honestly pumped and excited about it. I remember being superbly passionate about an idea and working hard to make it true and I can feel the similar energy. Why I am doing it?
The answer is to challenge myself. A lot of people including me believed that SAAS application were dead. The reality is far from over. SAAS revenues are higher and can be cracked if the right problem, right solution as a product and then the right creative messaging is executed properly. If you can crack any of the two, it makes it a great investible business, cracking all three gives you god like superpower of going direct to consumers. I feel what I am building with almondcoder - it is all three (for now).
What is AlmondCoder?
AlmondCoder is a programming tool similar to cursor, basically it helps you vibecode your projects easily. It has got many features which I think as a builder are really important such as,
- Run Claude Code / Codex / Cursor CLI in parallel, doing different things. No more waiting for a prompt to finish executing before you work on your next prompt.
- Create git subtree for each prompt and merge easily using an interactive GUI.
- Get a simple plan for your prompt which you can edit, thus giving you an overview of what changes the Claude Code is going to make and where to execute your prompt. Often at times claude code is not able to understand our prompts easily, devising a wrong plan and doing it all wrong.
- Having onboarding prompts (for open source projects, which allows others to discover your project architechture easily), the generic ones will help people to learn the architechture quickly.
- Prompt Pills, for each project you open we have allowed one to create pills which are recurring contexts appended to your prompt. You can define prompt pills for your project and select them.
- A lot of the software (prompt pills , onboarding prompts , auto merging) have features where some aspect of the software itslef is written on the fly based on users prompts given. This adaptability makes our software "create software on the fly".
We will eventually allow multiple people to work on prompt, plan, and get it approved before claude does it's magic.
Now the hard part to build this out is execution of such a tool is still little complex. This goes from designing the landing page, getting the right messaging, building out the product and pushing content across social media to get adoption. It is definitely not an easy task, especially as a bootstrapped founder (maybe easier as bootstrapped founder as the hunger to go to revenue is higher), but honestly speaking this is still very hard in the age of AI as well.
I will be documenting my journey in an structured manner on this blog if ever is considered in future. My todays' goal is to ensure that
> prompt creates a new git subtree.
> conversation history is being saved locally in a filedisk with specific json schema
> on clicking the items in prompt history I am able to load that specific prompt context.
> I am able to create new prompt with right prompts.
I will update it in builder journey - 2 how it's going. Obviously there are so many finer details as well assocaited with the above and need to make.
I will share progress so far.
Posts
