stablepeg

stablepeg

are the stablecoins still a dollar? for each coin: spot and a 10 minute twap from its deepest uniswap v3 pool against usdc, what a 100k swap returns on curve 3pool, and the coingecko price next to it for a sanity check. everything but the last column is read straight from ethereum mainnet over a public rpc. no keys, nothing to install.

$ stablepeg
stablecoins vs USDC on ethereum mainnet, block 25,929,333 (2026-09-08 01:22 utc)
USDC itself: $0.9999 on coingecko

coin    v3 pool       spot  twap 10m       dev  liquidity  coingecko
USDT    0.01%      0.99998   1.00000    -0.2bp    25.0e15     0.9999
DAI     0.01%      1.00001   1.00000    +0.1bp  9108.4e18     0.9998
USDe    0.01%      0.99997   0.99990    -0.3bp  4973.2e18     0.9997
FRAX    0.05%      0.99164   0.99164   -83.6bp    11.4e18     0.9921 !
LUSD    0.05%      1.00889   1.00884   +88.9bp    11.9e15     1.0090 !
PYUSD   0.01%      1.00004   1.00000    +0.4bp   709.8e12     0.9998
crvUSD  0.3%       1.00167   1.00170   +16.7bp   526.0e15     0.9999
GHO     0.01%      0.99902   0.99900    -9.8bp    54.6e18     0.9990
USDS    0.3%       0.99877   0.99880   -12.3bp     6.0e18     0.9998

curve 3pool: 100,000 USDT -> 99,987 USDC (-1.3bp), 100,000 DAI -> 99,985 USDC (-1.5bp), virtual price 1.0398

! = at least 50 bp off the peg: FRAX, LUSD

usdt, dai and usde inside one basis point is what "pegged" looks like. frax and lusd at 80-90 bp is not news, they trade like that most weeks; the flag is there so a real depeg would not hide in a wall of numbers.

install

pipx install git+https://github.com/alinaschanz/stablepeg

or clone it and run python -m stablepeg from the folder. python 3.10 or newer, no dependencies.

use

stablepeg                                       # the usual suspects against usdc
stablepeg --coins USDT,DAI,FDUSD,TUSD --quote USDT
stablepeg --twap 1800 --size 1000000            # 30 minute twap, 1m curve swap
stablepeg --warn 20                             # flag at 20 bp
stablepeg --depth                               # what a 100k / 1m / 10m sale returns, impact in bp (quoter v2)
stablepeg --json
stablepeg --no-coingecko --no-curve             # chain only
stablepeg --summary-append data/daily.csv --quiet   # one row per coin, the daily dataset

coins on offer: USDC, USDT, DAI, USDe, FRAX, LUSD, TUSD, PYUSD, crvUSD, GHO, USDS, FDUSD. the live test checks each address against symbol() and decimals() on chain, and the curve pool coin order against coins(i).

how it works

• uniswap v3 factory getPool(a, b, fee) for the 0.01%, 0.05%, 0.3% and 1% tiers; the pool with the most liquidity() wins. slot0().sqrtPriceX96 is the spot price: token1 per token0 = (sqrtPriceX96 / 2^96)^2 x 10^(dec0 - dec1), flipped when the coin is token1.
• observe([600, 0]) gives tick cumulatives; their difference over the window, floored, is the mean tick, and 1.0001^tick is the twap. pools with too few observations answer "OLD" and get n/a.
• curve 3pool get_dy(i, j, dx) for a 100k swap is an effective price with the fee in it, which is the number a real seller gets. get_virtual_price() is printed for the curious.
• coingecko simple price for the last column; skip it with --no-coingecko.
• every call is an eth_call over json-rpc via urllib; endpoints publicnode, drpc, mevblocker, tenderly, blastapi in that order, or --rpc yours. the fee tiers and the coins are fetched in parallel.

--depth asks the uniswap quoter v2 what selling 100k, 1m and 10m of the coin into the same pool would return; the impact is the effective price against the spot. a peg that holds for 100k and breaks at 10m is the interesting case. this is one uniswap pool: for dai and the newer coins most of the depth lives on curve and in the maker psm, so a big impact here reads as "not on uniswap", not as "depegged".

reading the table

• dev is spot minus one in basis points. ! marks at least --warn bp (default 50).
• liquidity is uniswap's L (sqrt of x times y in raw units). it is only comparable within a pair, so read it as "which tier the price came from", not as a ranking of coins.
• a 0.01% pool with real liquidity and a twap that agrees with the spot is a price you can trust. a 1% pool with a wide gap between spot and twap is a thin market, and the coingecko column will usually disagree with it.

the dataset

data/daily.csv gets one row per coin every night at 00:27 utc (spot, twap, pool, coingecko, curve): a small peg history that anyone can plot. columns are described in data/README.md; a rerun for the same day replaces the day.

exit codes and scripting

0 after a table, 2 when a coin is unknown or no rpc answered. --json carries everything the table shows plus the raw pool addresses and liquidity, so stablepeg --json | jq '.pools[] | select(.spot < 0.995)' is a depeg alert in one line; --summary-append is the same data as rows, one per coin per day.

see also

• bigmoves: the large transfers of these coins
• onchain-notes, gasweek, ens-lookup
• the notes: alinaschanz.life, the short version on x

verify a release

from the next release on, every release carries the sdist and the wheel, a SHA256SUMS file, an opentimestamps proof of that file, and a build provenance attestation made in github's own signing flow. with the files downloaded into one folder:

sha256sum -c SHA256SUMS
gh attestation verify ./*.whl --owner alinaschanz
ots verify SHA256SUMS.ots

the attestation names the commit and the workflow run that produced the file; the timestamp proves the checksums existed before a certain bitcoin block; the commit itself is signed.

license

mit. numbers, not calls. not financial advice.