AI-powered astrology API for developers. One endpoint, any question, expert answers. Vedic + Western astrology in one platform -Nvidia & Google Backed Firm

India
Build the experience. Plug in the astrology. Text answers. Turn-based voice replies. PDF reports. Vedika brings these product surfaces together through one API platform. Explore the keyless mock-data sandbox: vedika.io/sandbox
1
12
Give your customers an astrology report they can keep. Your logo. Your name. Your colours. Explore Vedika's English PDFs, embedded viewer and mock-data previews. White-label access is plan-dependent. vedika.io/pdf-reports
5
Test the gaps, not just timestamps. Five sampled times missed a brief chart state at about +3.45 to +6.32 min in a fictional test case. Calculation check only; forecast skill untested. Full paper: linkedin.com/feed/update/urn…
6
What happened between the checks? Five sampled times missed a brief chart state at about +3.45 to +6.32 min in a fictional test case. Calculation check only; forecast skill untested. Full paper: linkedin.com/feed/update/urn…
6
Full paper: linkedin.com/feed/update/urn… Five timestamps can miss what happens between them. In a fictional chart test, all five samples missed a brief chart state. We show why checking the gaps matters. A Birth Chart Is All You Need? A computational check, not a prediction test.
11
Your Death | Full paper: linkedin.com/feed/update/urn… What would a prediction of your death have to prove? We reconstructed a Ptolemaic timing rule. Its results are angular-time units, not ages. Historical review; no validated death-date prediction.
9
In Vedika’s synthetic experiment, a birth-time search produced 87% false positives under a deliberately flawed test; rerunning the search within the test gave 4.7%. For astrology and faith-app builders, how should a prediction earn trust? Our research series starts here.
1
15
The audit caught a separate issue: the artificial model’s expected future score was worse than the baseline even without a search. That was built into the model family. The search inflated historical fit; it did not explain that future-loss penalty.
1
9
A Birth Chart Is All You Need? is a research draft, not externally peer reviewed. It does not test production forecasts. More papers, including Your Death on longevity claims and mortality evidence, are coming before our November 1 launch. linkedin.com/feed/update/urn…
8
Google is rejecting scripting input on chrome and apple is also doing shit things with apple events ? So open ai how your computer use gonna make it ? I mean i want it to be developed further and faster but apple and google aint cooperating with open ai , seems so Fucking comeup with your own engine and dont let this tool die
21
I bet opus 5.5 is routing to opus 5 again The verbose and crap is back Now guys stop your work if u guys dont wana. Bug your system and create issues which later fable will fix and u guys will go mad that what a model
28
Did the model improve, or did the agent get another retry? We're building a more rigorous Vedika Bench and an enterprise harness with cloud VMs + multiple models. The goal: make tools, starting state, retries and outcomes easier to inspect. Both are in development.
11
New products from Vedika and Xalen arrive November 1. Over the coming weeks, we’ll share the demos, engineering decisions and evidence behind them, from astronomical computation to practical applications. Follow along for the first look.
Made with AI
1
24
@trq212 can u do something on computer use of claude code It sucks man and guard rails too , most of time it says i cant do this i cant do that where codex immediately goes in to using gmail sign in and driving it and certifying + acting on it where Claude comes sir u need to do
3
44
Mcp /connectors is bullshit when comes to dept work or real work
10
Even gpt bad models makes sense cos its visual tool and computer use is speed running things
9
Computer use tool - like if i have work in atlassian or jira where codex rocks and claude sucks and also for example if we have to do data models in crm or new setup , claude fails man
17
Opus 5.5 nerf has begain and already regressed 30% on task within day 2 lol Now medium feels low and extra feels medium And its already showing anthropic is pushing quantisation so its already fucked enough on day 2 and now guardrails fucking more up as claude is pushing hotfixes
1
62
Its already saying no to my own legit task and drifting too much and idk what verbose bullshit has came in back even despite consise mode is on and its speaking in confusing verbose also eaxtly how Claude devs post on X , i cant understand shit from opus 5.5 on verbose
1
25
Alert- Claude is back to opus 5 and lobotomised already Its uses less on day 2 itself and unable to use anymore and
2
1
21
Well claude is fucked already , their models only works on day 1 then 2 months night mare
4