[Contributor Proposal] EthicalHaus

[jilt]

Title: [Contributor Proposal] EthicalHaus — Transparent Financial Rails for Ethical Social Finance Trading

One-line summary

We are building a social-to-DeFi pipeline that transforms high-signal Farcaster interactions into real-time, ethical rewards using 1inch/hummingbot and Superfluid streaming.

Alignment with GoodDollar goals

This project addresses two primary ecosystem goals:

• Increase G utility by integrating $G streams as a primary reward token for social finance trading.

• Shift claimers→ earners: Providing a platform where high-quality creators earn constant G$ payment streams based on engagement rather than one-off claims.

3) Motivation & user value

• The Problem: Current SocialFi models are often speculative, extractive, and plagued by spam that buries high-quality content behind opaque algorithms.

• User Value: Our system empowers high-signal creators (Neynar Score > 80) and traders by providing transparent, gasless, and MEV-protected financial rails. Just as the Bauhaus school stripped away ornamentation for functional design, we provide “clean” infrastructure where the form of the system is dictated by its function: rewarding meaningful participation.

4) Scope & deliverables

• Deliverable 1: Integration of 1inch/hummingbot for gasless social-to-DeFi trades via Farcaster Mini Apps.

• Deliverable 2: Superfluid SPR module for G$ rewards, implementing a 5% cap on user flow rates and “no insider participation” rules to ensure fair distribution.

• Deliverable 3: Hybrid Data Layer (Postgres + Filecoin) to provide sub-10ms query speeds with a censorship-resistant cold archive.

• Deliverable 4: ERC-8004 Agent Identity registry on CELO to allow users to verify agent ethics and reputation before trading.

5) Technical approach

• Architecture: Farcaster Casts → Neynar Webhooks → Postgres/Filecoin Data Layer → 1inch/hummingbot Execution → Superfluid Streaming Rails

• Key SDKs: Superfluid SDK, 1inch Fusion+, Neynar API, and z.ai for API management.

• Risks: Balancing real-time stream updates with L2 gas costs.

6) Timeline & milestones

• Milestone 1: Launch Postgres/Filecoin hybrid data layer and Neynar agent. (Month 1)

• Milestone 2: Release 1inch/hummingbot Mini App and ERC-8004 agent identity minting. (Month 3)

• Milestone 3: Full Creatorcoin/Buildercoin G$ streaming integration and token launch with custom Uniswap v4 hook. (Month 6)

7) Testing & QA plan

• Functionality: Automated unit tests for Superfluid flow rate logic and manual QA within Farcaster/base app client environments.

• Security: Adherence to SIP #11 guardrails to prevent reward concentration.

8) Open-source & licenses

• License: MIT / GPL (compatible with GoodDollar and Superfluid standards).

• All used SDK are open source but 1inch, which is very important for MEV protection

9) Communication & progress updates

• Frequency: Weekly updates on the GoodDollar Discourse (KarmaGAP).

• Channel: Dedicated Telegram/Discord for real-time builder coordination on our API service.

10) Budget & payment expectations

• Requested Amount: 10,000 $ in G (calculated at current market rate).

• Purpose: To cover 6 months of development and farcaster social media curation operational costs, plus API subscriptions via z.ai, hosting, and Neynar Growth, estimated at ~$135/mo.

• Triggers:

  • 30% on completion of Milestone 1.

  • 40% on completion of Milestone 2.

  • 30% on completion of Milestone 3.

11) Links / references

• 1. Related Docs

• Governance: Aligned with Superfluid Season 5 SIP #11.

Team

_ Laura Camellini - fullstack dev

_ Agnes Jean-Baptiste - social media

[sirpy]

Hey.
Thanks for the proposal sounds very interesting. my feedback:

1. I had to dig into your deck to understand what exactly is the functionality of the app you want to build. I’d put that in the beginning and also here if you can share for everyone a one liner or a description of what your are building (ie whats the value for users?)
2. I’m not clear about how G$ will be used, can you share more specifics?
3. I’d like to see an alternative “reversed” milestones, meaning: First build a simple miniapp with minimal features, minimal complicated “decentraliztion” (postgres filecoin hybrid) etc. then start adding complex integrations like erc8004 and filecoin.

[jilt]

I am very sorry you had to make an effort to understand the project I am proposing, I will try to answer to each of your questions with a reference hoping this time the meaning will be clear.

What’s the value for users?

on slide 9 you can find the user flow,

In the first step you can see that all the flow starts with a cast, the cast doesn’t have to be about DeFi, if a user has a thicker connected to her/his farcaster profile and/or if active on a creatorcoin/buildecoin channel the bot can activate on that info.

In the second step the content quality is filtered based on the casted content (based on the info above) and the neynar score of the user (how plausible is that this user is a spammer).

In order to filter the content properly we need the neynar API paid subcription (75$/month, slide 10)

In the third step the erc 8004 agent steps in using a postgres db (official farcaster docs way of working with the protocol data) and historical data collected in our filecoin endpoint as a context to answer to the selected cast in a meaningful way, figuratively “chewing” the thicker/creatorcoin/buildercoin info through the 1inch/hummingbot for Celo routing and finding trade routes that may be interesting for the user who sent the cast.

The costs for this step are well explained in the slide 10, filecoin costs 5$/year so I decided not to include the cost in the list.

In step 4 the agent (z.ai subscription to open source LLM almost 4 times the amount of tokens of claude opus with the same investment) answers directly to the selected cast with a trading pathway meaningful for the casts’s context, using a mini app or maybe more appropriately using a snap, I have developed frames last year’s so I would probably find myself more comfortable building snaps (the new version of frames launched last week by neynar) but the result for the user is exactly the same, they and all their followers get a trading suggestion involving their affiliated tokens (creatorcoins or buildercoins) as a reply to their cast.

The cost of hummingbot/1inch API is listed in slide 10.

In step 5 user can review the trade and decide whether to opt in or live it be.

I hope I made the user flow slide clear and again I apologize for my lack of comprehensible language, focusing on how to build the infra sometimes makes the mind less articulate.

How $G will be used

The rewards from the trading (that will happen ONLY if the users choose to make the transaction using the funds in their farcaster/base app wallet) will be subject to fees that are built for our project to survive past the requested funding period, but given the fact that this is not a trading app we have added an option to STREAM (using the superfluid infra) the earnings from the trade in $G or $sup to the creatorcoin/buildecoin community, selected as the members of the creatorcoin channel (farcaster wallet) or the Liquid pool investors.

Minimal features

As I was saying to Lewis on telegram, without the neynar and 1inch API subscriptions there’s no sense in building this app, and I do not have the money to invest on them right now.

I have built for farcaster before for the $LUM community and two open source frames that allowed people to gate content using their NFT collections or their erc20 token, you can see them on my GitHub but, exactly and the objkt.com frames, they are not compatible anymore with the protocol and that is due to the fact that it keeps evolving.

The project is built to be autonomous after 6 months and to enable people to stream $G to their farcaster community with a few clicks, that was the point, I don’t understand why you should mention postgres and filecoin as the basic features as they are only infra, the idea’s basic features is social interaction.

[jilt]

furthermore I’d like to add that the required one line summary is the second sentence of my message, literally:

[jilt]

Also for a more articulated (maybe not enough for you) alignment with the $G utility integration I would step into the third sentence on my post:

I didn’t post only the slides, those are in depth explanations built to widen the message posted on this forum which took me hours to write.

[sirpy]

1. The flow of g$ rewards, g$ streaming etc is still not clear. trading of tokens can happen on multiple chains while G$ is only on specific chains. where do the rewards come from?
2. I understand 1inch can enable finding best trade routes, what is hummingbot for?
3. The one line summary doesn’t explain what is the actual functionality of the app. Which from what I understand is: the app replies to posts with an offer to buy a certain token (based on AI recommendations), users can then opt to buy the token by pressing the link which opens your miniapp. Is that a correct description?
4. You mention funding communities with G$ streams, I don’t understand how that ties to the app functionality.
5. When i’m talking about postgress/filecoin i’m not saying they are not needed. I’m also not saying that you shouldn’t buy subscriptions for AI/1inch. I’m saying your milestones first focus on complicated infrastructure, while you should and can first focus on building a POC with minimal infrastructure. For example you mention filecoin as a backup, nothing will happen if you build that backup in the last milestone. Integrating with erc8004 can also be delayed. Focus first on proving the social interaction works.

[jilt]

The 1inch Fusion API supports Celo.

https://medium.com/1inch-network/the-1inch-fusion-api-revving-up-the-performance-of-wallets-and-cexes-36d2c188ae40

Evidence from an ETHGlobal showcase project demonstrates that the 1inch Fusion+ Bridge implementation successfully supports Celo Alfajores testnet, allowing for multi-chain USDC swaps across five different blockchain networks, including Celo.

1inch Fusion+ Bridge | ETHGlobal

Additionally, the 1inch Network aggregates liquidity across 12+ blockchains, and Celo is among the supported networks for its cross-chain and swap functionalities

1inch and hummingbot are two alternative ways to achieve the very same goal, ensuring the stability of the service.

No, the app offers a trade, with multiple operations of buying and selling ensuring a profit in between, the profit is calculated and streamed, the fees that we take on the whole operations are calculated only on the profit, if there is no profit to be made by the trading the app will not propose a trading.

While accepting the trade the caster (who is selected using the neynar score for quality of content) can choose to keep the profit from the trade or to share it with the member of a channel or the contributors of a liquidity pool

Without the mev protection of 1inch the social trade interaction will fail every time (yes it covers the hummingbot swaps too), drastically and with money loss.

It was brought to my attention that many people here fo not know what a buildercoin or a creatorcoin is, I mentioned the example of $ralph to let people understand how not only the exist but fees from their transaction have been used for very good purposes already.

The idea of socialFi is the foundation of how farcaster gas grown in the past 3 months and I desperately hope that we will be able to steer its direction towards an ethically correct goal without tearing each other apart because we didn’t take the time to read the docs.

[sirpy]

looking at 1inch docs I dont see they support Celo.
I’ve sent you DM

[jilt]

yes the Celo chain is supported by the bridge in 1inch, as used and proved as working on Celo testnet by the second link in my answer to your question.