Accounts are created on an 'as requested' basis.
To request a new account or additional employee access,
contact tobias@tobystech.com.
After creating an account, remember to stay
in touch (complaints, questions,
comments, etc.).
Access is limited to building web applications
(i.e. Python and Rust server support is
not yet available).
However, in this stage you can get expert
software development and general support
for your application directly from the founder.
Contact tobias@tobystech.com
to apply for 'Implementation Support' (see more details below).
Applying for implementation support is currently the only way to
access the system.
"The future is built by passion and perseverance in the present."
Get inspired by some Video Tutorials.
By using the Development Web App, you agree to our terms of service.
The development web app system is a cloud-based development
environment. It allows developers to code freely, while benefiting
from built-in automations that assist in various parts of the
development process.
The system serves as a solid foundation on which to build
almost anything that can run on the cloud. This might be
as simple as a static website or as complex as a web app
with multiple UIs, databases, and
(not yet supported during 'Limited Access')
specialised processing servers.
Key principles that the development web app was built on
include simplicity (by eliminating repetitive tasks and
reducing surface area for bugs) and flexibility (by
allowing custom dependency installation and writing
standard code).
The development web app includes limited LLM code
generation tools that strictly augments developer
workflows.
The development web app can help developers explore complex ideas by creating personal projects, while spending more time learning or creating and less time struggling with setup and integration issues.
The development web app can be used to automate internal business processes (e.g. inventory management) or create tools for customers (e.g. booking platform) at a reduced cost. This reduced cost can make a previously unaffordable idea viable.
The development web app can simplify and speed up the development process and reduce maintenance burdens. This can result in more features being developed and to a higher standard, which would increase the core value-add and thus increase the competitiveness of the company. The system can be used for complex applications such as building AI agents.
(not yet supported during 'Limited Access')
The development web app can be used to: create UIs (e.g. for data recording and cleaning), conduct data analyses (including plotting live graphs or running advanced machine learning models), or other complex processes (e.g. running robotic simulations, running open source models, or training new models).
Managed environments can help developers focus on
strategic objectives rather than technical issues.
E.g. Improving the value-proposition could result in
10 times as many customers and in the context of
software this can mean 10 times as much profit, or more.
Focusing on setting up a custom environment likely won't
change the number of customers since customers don't
care about technical details as long as the product works.
If we exclude the cost of development time (both to
create and maintain a more complex setup), developers
may be able to reduce fixed costs, but this comes
at the increased risk of downtime, data loss, and
product launch delays.
Extra: Even in a less commercial context like passion
projects or internal tools, it is equally
beneficial to focus on strategic objectives rather
than re-inventing the wheel.
The main issue with managed environments is that they
can be very limiting, however the development web app
is a unique hybrid environment that is built first for
proper coding and only automates repetitive tasks.
The development web app was originally built to be an
internal tool to manage complex development projects.
It really is not a cheap tool for non-technical users.
Bonus: The development web app was built to not rely
on complex cloud infrastructure from major cloud providers.
This means that the output can run on smaller (cheaper
but still reputable) cloud providers.
We don't add markups
in this context, so we just choose the best cheapest
solution. While costs vary
dramatically, common alternatives can easily be 10 times
more expensive.
There is no hard vendor lock-in, the codebase (even code generated
by automations) is compiled in a
way that is legible and that follows industry standard practices.
The output is a Kubernetes deployment. This setup allows for
distributed computing to practically any scale.
The coding languages developers are expected to know include
JavaScript and SQL. In advanced contexts, developers may also
need to know Python (and in extremely advanced contexts
Rust).
Developers add their own API keys for computing resources and
other third parties. This
means that any resource usage is paid directly by the user without
any additional markup from the system.
Developers can connect
to any external third party using standard API calls. Third
parties could be anything from a SMTP mail server, an
external NoSQL database, or even some high-performance
computing resources.
For extremely custom processes, a separate manually managed
deployment can be made and then developers can connect to
that system with the system that they build on the
development web app. This may sound like more work, but at
the level of complexity where this becomes necessary this
splitting up is actually beneficial because it adds
modularity.
The system can be used for, but isn't ideal for, design focused static websites. The focus on flexibility does mean that technical skills are necessary. No-code tools may be better for simple projects, unless there are plans to increase backend complexity at some later date.
The development web app is a purely development tool. All management tasks (e.g. planning, task assignment and communications about strategic objectives) should be done externally, using generic external tools. Tools used in conjunction with the system can include third-party LLM tools or tools you built yourself using the development web app.
For certain problems that can be represented
effectively in training data, AI can find patterns
and infer ground truths. This can be useful for
finding novel solutions to these specific problems.
The mechanical nature of AI generally results
in much quicker problem solving in comparison to
humans. Which tasks
can be solved is limited to the training data
and the model architecture.
AI tools can be used in combination with humans to
capitalise on the relative advantages of either.
The timeframe and extent to which AI will replace tasks
is unknown. This is further complicated by the fact that
each task has its own set of challenges.
Additionally, we can speculate about limits about current
model architectures, but it's much harder to speculate about
what future breakthroughs might occur and when these might
occur. Even if a solution is available, timeframes are
further unclear because it is hard to determine how long
it will take for changes to be implemented into slow moving
industries.
Even if 'AI teaches AI', the model architecture used for this
cycle of learning can itself become the model
architecture that limits performance.
Additionally, models will still be limited by their training
data. We already have reinforcement learning, but not all
problems are easily defined in a way where we can generate
data that is suitable for this training.
The world is ultimately a complex place, creating better
model architectures, collecting better data, and applying
AI to new problems might be an indefinate source of work.
Embrace new technology instead of viewing it as a
threat, be flexible in order to adapt to changes, and
learn new skills in order to use modern technology
effectively.
Extra: Exercise caution and moderation. AI isn't
a black box that does not need to be understood or a tool
that is so powerful that it should be used for
everything. Instead use AI strategically for
specific problems while maintaining strict oversight
and most importantly a clear understanding of what
is going on.
1 employee
Free
Generic sales support only
Full access to the system
Less than 5 employees
€2 000 monthly fee
Developer support
Full access to the system
Less than 10 employees
€5 000 monthly fee
Developer support
Full access to the system
10 or more employees
€10 000 monthly fee
Priority developer support
Full access to the system
Regardless of your organisation's size, feel free to start on the free tier and work your way up as needed.
Monthly Fee: Negotiable
Beyond 'development support' an optional service is
'implementation support'. Implementation support
involves support beyond solving platform-specific
technical issues.
The idea behind implementation support is to provide
short-term access to talent in order to boost
organisational flexibility.
E.g. If an organisation wants to run an open source
LLM in order to not share sensitive data with a
proprietary LLM. Traditionally, if the organisation
does not have an employee with the skills to do this,
they would need to make the difficult decision of
determining whether it is worth hiring someone for
the project.
This decision is difficult if decision makers lack
expertise in the relevant field.
With implementation support however, if we have someone
with the required expertise, they can provide
a free consultation. Armed with a better understanding
of the problem and potentially new suggestions,
implementation support can then be used to initiate
the project. The person assigned to assist you will
create a minimum viable solution, that can be
implemented in a production setting.
Management can then review the project and decide
whether to add new or existing internal resources
to the project.
For maximum flexibility, there are no specific time frames
or rates. The requirements of the project will be discussed
and then we will offer a monthly rate to provide
implementation support. Amounts are prepaid at the start
of each monthly period in a subscription-style model.
The subscription can be canceled at any time — whether
the project concludes, priorities shift, or for any other
reason.
(a consideration)
At the moment, implementation support is limited to
the founder and thus the skillsets of the founder.
Contact tobias@tobystech.com to learn more.