The fact that the first version of Mimir is only scratching the surface of what we can do with an onchain oracle still baffles me. You can literally turn a LOT of onchain data into feeds that can then be used in Colada. Going long on staking TVL isn’t that far away. Perping everything.

The fact that the first version of Mimir is only scratching the surface of what we can do with an onchain oracle still baffles me. You can literally turn a LOT of onchain data into feeds that can then be used in Colada. Going long on staking TVL isn’t that far away. Perping everything.

Mimir now has a home. The website and docs are live. A first look at what we’ve been building for fully onchain, permissionless price feeds on Solana. Mimir Beta goes live September 29. mimirfeeds.io

Mimir now has a home. The website and docs are live. A first look at what we’ve been building for fully onchain, permissionless price feeds on Solana. Mimir Beta goes live September 29. mimirfeeds.io

Reading a pool price is easy. Making it expensive to manipulate is way harder. It has been the biggest challenge in building Mimir, and where much of our development time went. That work led to several safeguards built directly into how Mimir reads pools. An attacker cannot change a pool’s price and have Mimir read that pool in the same transaction. Mimir also checks its liquidity and the amount needed to move its price by 0.5%, rejecting sources below the feed’s requirements. Prices are combined using a weighted median: an attacker needs to influence sources carrying at least 50% of the total weight. Pools sharing the same liquidity route count once, not as separate confirmations. More independent sources can mean more markets to manipulate. Deeper liquidity means more money needed to move them. Prices too far from the median are excluded, and if too few sources remain, Mimir refuses the update. Mimir also estimates how much capital an attacker would need and how much they could lose attacking enough sources to move the price. From the capital estimate, it recommends a limit on the value an app should expose to that feed. Every check, limit and risk calculation is enforced onchain.

Reading a pool price is easy. Making it expensive to manipulate is way harder. It has been the biggest challenge in building Mimir, and where much of our development time went. That work led to several safeguards built directly into how Mimir reads pools. An attacker cannot change a pool’s price and have Mimir read that pool in the same transaction. Mimir also checks its liquidity and the amount needed to move its price by 0.5%, rejecting sources below the feed’s requirements. Prices are combined using a weighted median: an attacker needs to influence sources carrying at least 50% of the total weight. Pools sharing the same liquidity route count once, not as separate confirmations. More independent sources can mean more markets to manipulate. Deeper liquidity means more money needed to move them. Prices too far from the median are excluded, and if too few sources remain, Mimir refuses the update. Mimir also estimates how much capital an attacker would need and how much they could lose attacking enough sources to move the price. From the capital estimate, it recommends a limit on the value an app should expose to that feed. Every check, limit and risk calculation is enforced onchain.

First step is getting close. Mimir Beta will go live on September 29. On the Colada side, things are moving well. The web app is around 80% complete, and work on Colada Mobile has already started. The goal is to have Web and Mobile ready around the same time, so everyone can use Colada from day one. If everything goes as planned, we’re targeting Colada Mainnet Beta in October.

Mimir is introducing its own ticker registry. Each feed is represented by an NFT and gets a unique ticker of up to 7 characters, using letters and numbers. That ticker is how you find and access the feed directly in the terminal. While it’s registered to your feed, no other feed can use it anywhere in Mimir. Your feed. Its own name.

Quick update on Colada + Mimir. Mimir is on the final stretch before mainnet on September 29. You’ll get a terminal app to create feeds and track them in real time. At the same time, Colada Web is being tested in a mainnet sandbox against the latest version of Mimir. Work on Colada Mobile has also started properly and is already around 20% done. Mimir just went through its first deep wave of cyber audits too, with a few more expected before release. If everything goes smoothly with Mimir, Colada is likely next in October. Getting close.

09.29.26

Colada is clocking in 📱 We’re joining CLOCK IN, the Solana Mobile Hackathon by @solanamobile and @RadiantsDAO. Over the next few weeks, we’ll be building Colada for Seeker from the ground up. A mobile-first experience built around how people actually trade on their phones. The permissionless perp factory, built for Seeker. Clocked in until October 8.