As a developer I would like to price myself based on the price of my time and the time it takes to build these systems. The quality of work and speed would be top notch and the main issue could be client information dependency, scope realisation, and client coordination.
The timelines given below is considered with client coordination and iterations. Leaving external dependencies aside I assure that my development work wouldn’t exceeds the given timeline.
That being said we will always optimise on requirements and work on the tasks without waiting on for information from the client.
So I’m setting a price of $6,000 / month for any project that we bring in initially.
This price is the exact match of the previous job and I want to keep the base price for the first project that we work together.
Feel free to markup and charge whatever you want for this project, I don’t mind ;)
All the infra costs should be beared by the client as it can shoot up based on usage
The infra costs for base usage are minimal for non AI stuff vercel 20 USD / m, supabase 20 USD / m video/image data, twillio per call, per message, per email pricing,
AI APIs get expensive if the client demands SOTA models, but we can do good with mini / lite models to keep infra costs low. These days lite models are 10x more better than GPT 3.x, 4.x versions as well ~2 years old…
Maintenance cost would be $400 / month where I manage apis, rate limits, scaling compute to optimise infrastructure costs, keep backups to keep the system safe and any other infra issues like api changes, data corruption and more
This is the expected time it would take me to build these features with detailing and iterations with the client. Some features are easier done in sequence and are dependent on each other. I have given the timeline assuming that we take the optimal timeline of tasks that stack one on another