Update December 8, 2024:
- Added table of contents
- Better pagination
- More budget details on page 6
- Developer outreach on Page 9
----------------------------------------------------------------------------------------------------------------------
Update December 5, 2024:
On page 7 of the document, we added a part called "AI architecture" to explain our architectural approach better.
-----------------------------------------------------------------------------------------------------------------------
Hello everyone!
This discussion post aims to gather feedback from the community about our upcoming treasury proposal for dAppForge. We are looking forward to receiving feedback!
1. Project
1.1 Summary of Important Points
We submitted a previous treasury proposal, ref 985, and completed all the tasks/milestones.
Milestone 2 of dAppForge V1 is now completed. Here are some videos that show what you can do with the plug-in:
Here is our GitHub: https://github.com/dappforge
Currently, the plug-in can be used for code generation and autocompletion for substrate and ink! The plug-in does a good job of providing insights into the code generated, which helps developers increase their understanding of substrate/ink. With every fine-tuned version, developers should be able to build/code better and faster. This will lead to more applications being built on Polkadot.
We have now completed 2 of the 4 milestones to build dAppForge V 1.0. To recap, these are the 4 milestones:
- Milestone 1 (completed): the first iteration of LLM ( substrate only) and the VS code plugin alpha version. This was completed, but the product is at a level where developers won’t see a lot of value. This is because the model needs to be further fine-tuned.
- Milestone 2 (completed: ref 985): The product's beta version. It includes a fine-tuned LLM model for substrate, integrated ink!, and a chat feature.
- Milestone 3 ( current treasury proposal): These are the main points ( more details in section 1.6):
- Fine-tune llm for substrate and ink! further. This will help developers increase their understanding of both substrate/ink!, which will lead them to code faster and reduce development time.
- Integrate Solidity. This will set the foundation for helping developers deploy EVM smart contracts on Asset Hub when Plazza is finished in Q2 2025. Polkadot needs more dApps, and one strategy is to make it easier for solidity developers to deploy on Asset Hub.
- Develop and integrate a monetization strategy to support the product's growth, focusing on a two-tier pricing structure, initial freemium experience, and scalable subscription option ($30/month). The paid tier will give developers access to enhanced AI, providing more accurate and contextually relevant code suggestions. We will also build a specialized educational subscription track focused on onboarding users to learn using Substrate and Ink!
-
dAppForge is open-sourced. Anyone can clone the AI model and run it on their machines. However, you will need to set up your cloud infrastructure to run the model. If you don’t want to manage your own infrastructure, there will be a free tier for you to use.
- Develop a knowledge graph focused on the Rust programming language, capturing comprehensive insights on Rust syntax, libraries, frameworks, best practices, and common patterns. This will be crucial to expanding our reach to the wider Rust developer community. It will also create an opportunity to onboard them into the Polkadot ecosystem by demonstrating how their existing skills can be used to build blockchain applications using substrate/ink! Here are some examples:
- Upon starting a new project, the plug-in offers a dropdown option: "Would you like to try building a decentralized application using Ink!?" or “Would you like to explore how to build your blockchain using substrate?”
- Once JAM is live, we expect Rust developers to write contracts in Rust that would be compiled with PVM. We also expect rust to be used to write different JAM services in the future.
- Milestone 4 ( future treasury proposal): These are the main points:
- Fine-tune current knowledge graphs.
- Enable Ink! and Solidity developers to deploy their smart contracts to Asset Hub. This involves making sure the contracts correctly compile to PVM.
- Implement more features for the paid tiers, such as:
- Personalized Learning: Train the AI on your codebase for tailored suggestions.
- Up-to-date Knowledge Graph: Access the latest Polkadot knowledge and best practices with our regularly updated knowledge graph.
1.2 Overview
Tagline: AI-powered plug-in for Polkadot developers that reduces development time.
Brief description: Building on Polkadot is complex, creating a high barrier to entry. dAppForge aims to enable Polkadot developers to build better dApps, appchains, and parachains faster, lowering the entry barriers for new developers. The AI model is currently trained on Substrate and ink!. The model can used for code generation and autocompletion for substrate and ink! It provides insights into the code generated, which helps developers increase their understanding of substrate/ink. We will now also integrate solidity to enable developers to deploy EVM smart contracts on Asset Hub when Plazza is completed in Q2 2025.
Motivation: Substrate and ink! can build great parachains, appchains, and dApps, attracting millions of users, which is crucial for the blockchain industry. Polkadot has the second-largest developer ecosystem and consistent growth in developer activity. dAppForge aims to empower existing developers and make it easier for new developers to join and build on Polkadot. Additionally, including solidity will make it even easier for developers to deploy their smart contracts on Asset Hub.
Here is the rest of the proposal: