Resources.
All
Case Study
Experiment
Opinion
Reduce your GPT-5.6 API spend by matching model tiers to workload.
We stopped treating Sol as a universal default. By routing routine tasks to Terra and Luna, we reduced spend while keeping complex agents on the right tier.
How we cut 90% of our AI token costs by running Ollama on an old RTX 3090.
We were using cloud LLMs for Excel data cleaning and it was becoming unpredictable and costly. By building our own machine we eliminated recurring token fees, secured client data, improved privacy and reduced monthly AI spend from hundreds of pounds to the cost of electricity.
We saved an accountancy practice over 30 hours per month by optimising a single internal process using AI.
Code named project Orion we optimised an internal process helping cut the time taken to perform it by 75% saving our client over 30 hours of time per month.
Changing your Wordpress page builder — how hard can it be?
Learn how we migrated one of our clients from Divi to Elementor. A two week process that was way more complicated than we first thought.
All
Case Study
Experiment
Opinion
Reduce your GPT-5.6 API spend by matching model tiers to workload.
We stopped treating Sol as a universal default. By routing routine tasks to Terra and Luna, we reduced spend while keeping complex agents on the right tier.
How we cut 90% of our AI token costs by running Ollama on an old RTX 3090.
We were using cloud LLMs for Excel data cleaning and it was becoming unpredictable and costly. By building our own machine we eliminated recurring token fees, secured client data, improved privacy and reduced monthly AI spend from hundreds of pounds to the cost of electricity.
We saved an accountancy practice over 30 hours per month by optimising a single internal process using AI.
Code named project Orion we optimised an internal process helping cut the time taken to perform it by 75% saving our client over 30 hours of time per month.
Changing your Wordpress page builder — how hard can it be?
Learn how we migrated one of our clients from Divi to Elementor. A two week process that was way more complicated than we first thought.
All
Case Study
Experiment
Opinion
Reduce your GPT-5.6 API spend by matching model tiers to workload.
We stopped treating Sol as a universal default. By routing routine tasks to Terra and Luna, we reduced spend while keeping complex agents on the right tier.
How we cut 90% of our AI token costs by running Ollama on an old RTX 3090.
We were using cloud LLMs for Excel data cleaning and it was becoming unpredictable and costly. By building our own machine we eliminated recurring token fees, secured client data, improved privacy and reduced monthly AI spend from hundreds of pounds to the cost of electricity.
We saved an accountancy practice over 30 hours per month by optimising a single internal process using AI.
Code named project Orion we optimised an internal process helping cut the time taken to perform it by 75% saving our client over 30 hours of time per month.
Changing your Wordpress page builder — how hard can it be?
Learn how we migrated one of our clients from Divi to Elementor. A two week process that was way more complicated than we first thought.
Let's talk — email hello@heavymask.com
© 2024 This site and its contents are owned by Heavy Mask, UK registered limited company. For All inquiries, please refer to our full terms and conditions.
Let's talk — email hello@heavymask.com