adriatic / benandnik Goto Github PK
View Code? Open in Web Editor NEWrepository for communications between Ben Lower and Nik Ivancic
License: MIT License
repository for communications between Ben Lower and Nik Ivancic
License: MIT License
I do not believe that I changed any code - and it ran yesterday
npx fixie@latest serve
🦊 Serving agent fixie-sidekick-template...
🌱 Building agent at /Users/nik/dev/work/fixie/sidekick_tutorial...
npm notice
npm notice 📦 [email protected]
npm notice === Tarball Contents ===
npm notice 1.1kB LICENSE
npm notice 512B README.md
npm notice 86B dist/index.d.ts
npm notice 1.8kB dist/index.js
npm notice 95B dist/system-message.d.ts
npm notice 3.2kB dist/system-message.js
npm notice 637B package.json
npm notice === Tarball Details ===
npm notice name: fixie-sidekick-template
npm notice version: 0.1.0
npm notice filename: fixie-sidekick-template-0.1.0.tgz
npm notice package size: 3.5 kB
npm notice unpacked size: 7.5 kB
npm notice shasum: 6c313edf8f150a80bf11354ee6edde48238fdeed
npm notice integrity: sha512-6AsLo8ROmmxAM[...]gchtOwBsfmJ/A==
npm notice total files: 7
npm notice
fixie-sidekick-template-0.1.0.tgz
🌱 Running: npx --package=@fixieai/sdk fixie-serve-bin --packagePath ./dist/index.js --port 8181
🌱 Agent process running at PID: 83138
🌱 Watching /Users/nik/dev/work/fixie/sidekick_tutorial for changes...
🚇 Starting tunnel process...
🌱 Agent child process started with PID [83138]
🥡 Replacing current agent revision f670933b
🌱 Agent stdout: file:///Users/nik/dev/work/fixie/sidekick_tutorial/node_modules/@fixieai/sdk/dist/instrument.js:70
globalThis.navigator = {};
^
TypeError: Cannot set property navigator of #<Object> which has only a getter
at file:///Users/nik/dev/work/fixie/sidekick_tutorial/node_modules/@fixieai/sdk/dist/instrument.js:70:22
at ModuleJob.run (node:internal/modules/esm/module_job:218:25)
at async ModuleLoader.import (node:internal/modules/esm/loader:323:24)
at async loadESM (node:internal/process/esm_loader:28:7)
at async handleMainPromise (node:internal/modules/run_main:120:12)
Node.js v21.6.0
🌱 Agent child process [83138] exited with code 1
Note these images are critically important to anyone trying to create corpora and is a LLM programming newbie - as the second one is used in the process of creating "Git&Github corpus"
and
Let's use Ben Lower (as a normal - not privileged user) to verify this initial test. I created a replica of your agent creation defined in Youtube presentation https://www.youtube.com/watch?v=Q5-spoD7Tng.
The url of the agent is https://embed.fixie.ai/agents/7b89bdfb-c948-4879-8300-b6ede770f04c?debug=1
Sharing this URL with a friend will not work as it will prompt that friend to login to fixie.ai
HTML to embed this agent (smallest fixie app) is
<!-- Embeds the agent chat UI into an existing page. -->
<iframe src="https://embed.fixie.ai/agents/7b89bdfb-c948-4879-8300-b6ede770f04c?debug=1" allow="clipboard-write" />
If I create a minimal RedwoodJS app and deploy it at say, Netlify, giving the URL of that deployed app would still prompt for Fixie login.
Clearly we need to discuss it further - it seems that the best solution would be to provide a class of agent IDs that would allow running such app (agent) without specific authentication
Note: I do care about people's perceptions on fixie documentation - and typos like these two do not pass my muster.
It is this link
as well as this one
Point to https://docs.fixie.ai/
This is really an offensive problem, as every user will trip there 😸
Fixie.ai is an innovative platform designed to simplify the integration and application of conversational AI and large language models (LLMs) in various business and development contexts. Founded by former engineering leaders from Apple and Google, Fixie.ai has quickly positioned itself as a significant player in the rapidly evolving field of generative AI, particularly in creating applications that leverage the capabilities of language models similar to OpenAI's ChatGPT
2
.
AI.JSX Framework
One of the standout offerings from Fixie.ai is the AI.JSX framework, an open-source tool designed to be intuitive for developers familiar with React and JSX. This framework facilitates the easy creation of UI components, DocsQA, and the integration of tools to enable generative UI (GenUI) through conversational AI. AI.JSX aims to bridge the gap between the extensive Python ecosystem used in AI and ML development and the JavaScript/React ecosystem, which is pivotal for creating leading end-user experiences
1
.
Enterprise Platform and Funding
Fixie.ai has raised $17 million in venture capital funding, underscoring the industry's confidence in its approach to making it easier for companies to build on top of language models. The platform is described as enterprise-oriented, offering an infinitely extensible model that allows businesses to integrate LLM-powered agents into their products and services. These agents can be built by anyone and run anywhere, providing a flexible solution for enhancing enterprise data, systems, and workflows with AI capabilities
2
.
Pricing and Plans
The platform offers a range of pricing plans to accommodate different needs, from developers and small projects to serious businesses and custom enterprise solutions. Features vary across plans, including the number of LLM tokens per month, support for document crawling and ingestion, security features, and support options
3
.
Multi-Modal Capabilities and Sidekicks
Fixie.ai is designed to be multi-modal, handling not just text but also images, video, and more. It allows for the creation of "Sidekicks," which are conversational AI agents that can be customized to perform a wide range of actions, from fetching real-time information to calling APIs. These Sidekicks can be embedded directly into applications, providing a seamless integration of AI capabilities
5
.
Open Source and Community Engagement
Fixie.ai emphasizes community engagement and open-source development. The AI.JSX framework is open-sourced under an MIT license, encouraging developers to contribute to and extend its capabilities. Fixie.ai's approach to building and hosting smart agents that extend the power of LLMs is indicative of its commitment to advancing the field of conversational AI through collaboration and innovation
4
6
.
Conclusion
In summary, Fixie.ai represents a significant advancement in the application of conversational AI and large language models within the enterprise and development communities. Its AI.JSX framework, enterprise platform, flexible pricing plans, and commitment to open-source development and multi-modal capabilities position it as a versatile and powerful tool for businesses and developers looking to leverage the latest in AI technology.
The prompt for both responses is the same: explain adding custom form validators
Adding custom form validators in RedwoodJS involves leveraging the validation
prop available on form input components like TextField
, TextAreaField
, etc., which accepts all of React Hook Form's register options. This prop allows you to specify custom validation logic directly within your form fields, making it possible to enforce rules such as required fields, pattern matching, or custom validation functions.
For instance, when you're creating a form field with RedwoodJS, you can specify validation rules directly in the component. Here's a simplified example to illustrate how you might add validation for a required email
field with a specific pattern:
<TextField
name="email"
validation={{
required: true,
pattern: {
value: /[^@]+@[^\.]+\..+/,
message: "Please enter a valid email address",
},
}}
/>
In this example, the TextField
component for the email address is marked as required and must match a regular expression pattern that validates the format of an email address. If the input does not meet these criteria, the form will display an error, helping ensure that users submit valid data.
RedwoodJS forms are designed to simplify the process of handling form validation, abstracting much of the complexity involved in managing form state and validation logic. By utilizing the validation
prop, developers can define a wide range of validation requirements directly within their form fields, making it easier to ensure that the data collected through forms is accurate and meets the application's requirements.
For more complex validation logic or to understand the full capabilities of RedwoodJS forms and React Hook Form's validation options, refer to the official RedwoodJS documentation on forms and React Hook Form's documentation【6†source】【7†source】.
I got an immediate feedback from RedwoodJS:
"Just to let you know, we are already using https://orbit.love/community-search and Orbit to something similar and they will soon have AI features and also plan to have bots for Discord and Slack etc."
which I have not yet shared with Anir&Co, because the additional comment I got from RW:
"By using I mean we’re trialing it. We haven’t decided whether or not to use in in Discord"
I responded back with:
The key feature I would like to point out is that collected data obtained from analyzing Discord will be feed that into RedwoodJS LLM, resulting with a full circle of continuously improving RedwoodJS.
Note that this product could become one of the future Fixie tools if your team decides that "just making Fixie" as it is today may not be sufficient differentiator from everyone else that jumped in AI. That is why I am not sharing this with Anir yet.
I just created the RedwoodJS corpus and related agent
Fixie Corpus | Fixie Agent |
---|---|
I did not go any further, since it is very easy to verify the functioning of this sidekick by pasting it's URL (see below) into the browser of your choice (except of Safari, where the agent does not work - issue for Fixie support team).
https://embed.fixie.ai/agents/ed9fa76c-a94b-42ba-81af-5470c3329ade?debug=1
Real test from a RedwoodJS core team member, who advised Keith Elliot that his sidekick needs improvements to correctly respond to a prompt like:
tell me how to add custom form validators
Here is what I got from my agent:
I am trying to build the "sidekick-github" from scratch (actully starting from https://github.com/fixie-ai/fixie-sidekick-template/tree/main).
nvm use 18
)npm run dev
yields the following error (Problems tab in VSCode terminal)File '@tsconfig/node18/tsconfig.json' not found.
The same issue as rendered on the VSCode editor:
Chrome Version 121.0.6167.139 (Official Build) (x86_64), macOS Sonoma, Version 14.2.1
While running the demo https://www.youtube.com/watch?v=Q5-spoD7Tng, I observed the following
To better illustrate this idea, I am sharing the key moments from a similar idea created by Microsoft - GitHub Copilot in VS Code. This feature is a lot more generic as it is applicable to any software development that utilizes Github in its workflow. Contrast this with the Fixie VScode extension, created using RWJS corpus - which makes this extension assisting RedwoodJS developers.
The next improvement would be to create as many Fixie VSCode extensions as there are application development frameworks. In the process of creating a first few, I expect the ability to optimize that by making the Fixie VSCode extension that supports many frameworks - and possible make the decision on any given framework dynamically.
A declarative, efficient, and flexible JavaScript library for building user interfaces.
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
An Open Source Machine Learning Framework for Everyone
The Web framework for perfectionists with deadlines.
A PHP framework for web artisans
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
Some thing interesting about web. New door for the world.
A server is a program made to process requests and deliver data to clients.
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Some thing interesting about visualization, use data art
Some thing interesting about game, make everyone happy.
We are working to build community through open source technology. NB: members must have two-factor auth.
Open source projects and samples from Microsoft.
Google ❤️ Open Source for everyone.
Alibaba Open Source for everyone
Data-Driven Documents codes.
China tencent open source team.