oh! I am just recalling an issue someone else had which I think might be related to this! Will fix it π«‘
In the meantime - here is a quick video for changing sessions.
If it happens again - can you try changing sessions like in the video below and lmk if the issue persists?
Thank you so much for letting me know about this!
(i think I know the underlying cause - so hopefully a fix will be shipped before you have to use this tip)
This is one of my favorite cvforge uses - ask it for a way to express a trade idea! Example:
prompt: I want to buy-the-dip on $AVGO which just fell after earnings to below 420.
I think it will recover to 500 by EOY.
What is the most capital-efficient, and risk-reducing way of achieving this?
It built a "lab" to compare the best ways to express this trade!
Watch cvforge build it live πͺ
It suggested the 420/500 call spread for the following reasons:
capital-efficient:
selling the 500 call pays for ~β of the 420 call, so you control the 420β500 move for ~$35.6/share instead of ~$50 (naked call) or $416 (stock).
risk-reduced:
max loss is the debit, full stop, and the short leg cancels most of the long call's theta and vega, so an IV crush after the earnings spike barely touches you (drag the IV slider down to ~0.7Γ and watch how little the spread moves vs. the naked call).
thesis-matched:
your edge is "500 by EOY," not "600." Giving up gains above 500 is exactly what makes it cheap. For equal risk it pays more at 500 than the stock or the naked call.
But is also let me compare with:
- long call
- put spread
- long stock
cvforge is free! π
interesting!
I have been alternating between both. They both seem to work perfectly fine - the difference is in aesthetic and also how much they go "above and beyond" what I've asked them. codex seems to really double/triple check everything.
the continuos refresh I've seen it too - when it is making a lot of edits, but then it finishes and it works fine.
What you mention regarding not being able to type anything on the prompt is weird tho - that shouldn't happen. Questions:
>are you on windows or mac?
>how often has it happened?
>does it happen when the agent has finished and you cant type or while it is working?
Thank you for your comments!
@JackFelke Awesome! I hadnβt thought about it being OSS but that might be very cool! I love the idea of people contributing - let me ping you back on this, thank you for asking!
cvforge glues all the pieces together so you have a single step between "prompt" -> "options analytics".
And its free! (includes options data)
see it live πͺ
prompt: "Give me a visualization to compare the options flow of $NOK and $IREN."
Give it a shot! lmk how it goesπ
@OhHeyAlan can you ask it to look at I:SPX directly (like the prompt suggests)?
I just tried the following prompt right now and it worked "give me a bar chart of volume for I:SPX"
The data is definitely there - so maybe a fix on the prompt/code should work
prompt: Create a scanner for unusual OTM options in non-ETF tickers.
I want to be able to filter by: delta, volume/oi ratio, volume, oi - as well as sort.
Lets have a fire emoji (1 to 3) to identify the severity of the unusualness.
watch cvforge building it liveπ«‘
It immediately identified some $NVDA and $MRVL activity!
Here is the agent reasoning verbatim:
βΊ Let me look at the available fields and sample some
data to design around.
βΊ This dataset has no ETF/asset-class flag, and the
universe includes thousands of ETFs (broad-index,
sector, and single-stock leveraged ones like
AAPU/AAPD). My plan: pull the most-active contracts
via screen, then client-side compute OTM status +
vol/OI ratio, filter out ETFs using a curated denylist
(plus an editable exclusion box), and score
unusualness with π₯βπ₯π₯π₯. Let me verify the
OTM/ratio logic against real data before building.
βΊ Logic confirmed β OTM calls show strike > underlying
with small positive delta, and I can see genuinely
unusual names ( $NVDA $245 call at 6Γ vol/OI, $MRVL $320
call). Now I'll build the scanner.
The scanner is built. Here's what it does:
Filters (all live controls):
βΊ The scanner is built. Here's what it does:
Filters (all live controls):
Filters (all live controls):
- Side β All / Calls / Puts
- Min/Max |Delta| β defaults to a 0.05β0.40 band (OTM
territory; <0.50 = out of the money)
- Min Vol/OI β the unusualness threshold (β₯1 = volume
exceeds existing open interest)
- Min Volume and Min OI β liquidity floors
- Exclude tickers β free-text box to knock out any ETF
stragglers
Sorting β click any column header to sort (toggles
asc/desc); defaults to the unusualness score
descending.
π₯ severity β driven by the volume/open-interest
ratio:
- π₯π₯ β₯ 1Γ
- π₯π₯π₯ β₯52.he
- π₯π₯π₯ing β₯ 5Γ (a big multiple of fresh volume over
existing OI = new, unusual positioning)
Non-ETF handling β this is the one judgment call worth
flagging: the dataset has no asset-class flag, so I
can't query "stocks only" directly. I handle it two
ways:
1. I:-prefixed index products (SPX, VIX, etc.) are
auto-excluded.
2. A curated denylist of ~600 well-known ETFs/ETNs
(broad index, sector, leveraged, bond, commodity,
single-stock-leveraged like TSLL/NVDL/AAPU, income
funds) is filtered out client-side.
That covers the liquid ETFs that would otherwise
dominate an unusual-volume scan, but it can't be
exhaustive across the thousands of ETFs in the
universe β so the Exclude tickers box lets you knock
out anything that slips through.
How it works under the hood: it pulls the 800
most-active contracts above your volume/OI floors via
screen, then computes true OTM status (call strike >
spot, put strike < spot), the delta band, and vol/OI
ratio client-side (so puts with negative delta work
correctly).
It auto-runs an initial scan on load. Can you see the
results table? A couple of things I can tune next if
you want:
- If real-but-obscure ETFs keep appearing, I can grow
the denylist or add a suffix heuristic.
- I could add a % OTM range filter, or factor absolute
volume more heavily into the fire score.
cvforge is π₯
06-03 Daily Summary
Welcome to the summer doldrums! RV compression in full effect; approaching levels soon that have HISTORICALLY seen a reaction.
Market Data from @ConvexValue
-Volatility expectations (Term structure) moving up slightly from the lows; yellow = today, red = yesterday
-GX Surface from yesterday shows a more balanced market with gamma peaks up at 7680, gamma flip near 7400
-Overnight straddle is ~30 points giving us an ON straddle range of [7580, 7640]
-Weekly Straddle ~90 pts gives us [7490, 7670]
-Overnight Volumes are average - slightly more puts; 7650 calls, 7600 puts highest volume
Net Exposure (SPX) from @OptionsDepth
-Pivot is 7600
-Targets above are 7615, 7625, 7635, 7650
-Targets below are 7580, 7550, 7525, 7510
Thoughts
-7637 looks like an interesting close but let's see
prompt: I think $MRVL will go up to $500 within the next 6 months, but it can also crash.
I am willing to risk a crash to $200, but not more.
Find me the best trade for this.
cvforge builds it live π It compares call/stock/stock+put structures.
The thing about creating your own software with cvforge - you can ask anything!
You are not restricted to whatever the off-the-shelf analysis is. You can create your own analytics and your own sites to answer one-off questions.
This dashboard here is a one-off. I might ask it again if I consider the trade again, but I might off.
It gave me an answer and thats all I neede from it!
cvforge is free (includes options data)!