Daniel - Bountycaster pfp
Daniel - Bountycaster
@pirosb3
Mini App devs: What’s the #1 pain when going live? Which boilerplate should just disappear? What is stopping your from shipping your first app? Drop your pain points below
8 replies
5 recasts
24 reactions

wasyl pfp
wasyl
@wasyl
I would add traffic, pressuring infra. If your miniapp gets a few shares and has some token, nft, 'value' related abilities - the amount of spam and farmers that roll through can quickly put pressure on any services like, Vercel hosting, Neynar API rate limits, etc Yes you can check spam scores, but this is still an API call and typically apps have 'other api calls' happening instead. Having to upgrade your infra and services just to handle low quality users.
1 reply
0 recast
2 reactions

wasyl pfp
wasyl
@wasyl
For small time miniapp devs, are they ready to handle the traffic from farmers and bots that provide no value to their app? Then most of your dev and time gets spent preventing or reducing those users. Most of your bill goes toward handling all that low quality traffic
1 reply
0 recast
0 reaction

Daniel - Bountycaster pfp
Daniel - Bountycaster
@pirosb3
Really helpful insight - thanks! Which signals would let you flag high-value users vs. bot/farming traffic? And once you have that data, how would you adjust your mini app (rate limits, rewards, UX, etc.) to act on it?
1 reply
0 recast
0 reaction

wasyl pfp
wasyl
@wasyl
Checking neynar user score helps, BUT as we've seen: 'new users' start at 0.5 ( which our miniapp targets ) and there have been a few farmers/spammers that show higher than 0.5, even above 0.6. We flag users that show signs of spam, claiming airdrops quickly, performing actions suspiciously quickly, and doing peer reviews of red flag profiles. If they have a lot of copy-pasta, share/claim crap ( no original content ), it's a red flag. The amount of users we have to block, ban, or prevent is an absurd amount, you would not believe the number. Users that dont provide value, try to take advantage, and only increase our infra bills
4 replies
0 recast
1 reaction

Valerie pfp
Valerie
@vrose.eth
Can you tell us more about "doing peer reviews of red flag profiles"? What is the type of peer review and what makes a profile red flag?
1 reply
0 recast
1 reaction

Valerie pfp
Valerie
@vrose.eth
Can you share some examples of spammy accounts with scores above 0.5 or 0.6, plz? We can use that data to keep improving the algo.
1 reply
0 recast
1 reaction

rish pfp
rish
@rish
Why don’t you set it to 0.65?
1 reply
0 recast
1 reaction

Neynar pfp
Neynar
@neynar
manually looked through a lot of these. what we found: - we can reduce the amount of time users have a neutral 0.5 score, this is in place now - for users around 0.6 (over/under), they are mini app players who mostly engage with Farcaster by playing and sharing mini apps. They used to get lower scores before but given that this is a genuine case case, we actually made a change that allows them to move slightly above 0.5 if all they are doing is playing mini apps and not spamming other people. More on that here: https://neynar.com/blog/retraining-neynar-user-score-algorithm - there was 1 user who was sneakily using bots to increase their scores (rare), this will get fixed
0 reply
0 recast
0 reaction