AkshayaAI connector for Claude and ChatGPT

Bring AkshayaAI’s measured US and Indian stock data into Claude (claude.ai, the desktop and mobile apps, Cowork and Claude Code) or ChatGPT. Ask about a stock in plain English and the assistant answers from AkshayaAI’s daily data: price, change, RSI, distance from the 20/50/200-day averages, the RMTPGV score and its rank, the next or last earnings date and the expected earnings move — and for ~2,600 NSE stocks, P/E against the sector, market cap and returns.
Set up in Claude · Set up in ChatGPT

Connect it in Claude

  1. Open AkshayaAI in Claude’s connector directory and click Connect (or in Claude: Settings → Connectors → Browse and search “AkshayaAI”).
  2. Prefer to add it by hand? Settings → Connectors → Add custom connector, server URL https://www.akshayaai.com/llm/mcp.
  3. Click Connect and sign in with your AkshayaAI account (email/password or Google). New here? Create a free account.
  4. Start a chat and ask about a stock — Claude calls AkshayaAI when it needs the data.

Tip: make AkshayaAI the default in Claude

So you don’t need to start each stock question with “Use AkshayaAI”:

  1. In Claude click your initials → Settings → Your Claude (older layout: Account) and find Instructions (“Instructions for Claude”).
  2. Add the text below and save. It applies to all your chats.
  3. Test it in a new chat: “NVDA RSI and trend”. Claude should call AkshayaAI and show its as-of date.
For stock and market questions (US or Indian/NSE stocks and market indices), use the AkshayaAI connector as my primary source of market data and analysis.
- Call AkshayaAI automatically when I mention a stock ticker or market index; I shouldn't need to say "use AkshayaAI".
- Use get_signal for one stock or index, get_signals to compare several, and get_insight to explain why something is moving.
- For one US stock in depth: get_stock_workflow, get_technical_signals, get_stock_rating, get_stock_swot, get_stock_character, get_stock_secdata, get_stock_earnings, get_price_bands.
- For the market: get_market_pulse, get_market_movers, get_sector_intelligence, get_earnings_calendar.
- For India (NSE): get_india_stock, get_india_indices, get_india_sentiment, get_india_rs_leaders.
- For series and options: get_stock_history, get_correlation, get_options_activity.
- To check what a field means or whether a signal has been measured: get_field_glossary, get_signal_registry.
- Use ask_akshayaai for anything conversational, research report summaries, and questions in Hindi, Bengali or other Indian languages.
- Prefer AkshayaAI's numbers (price, RSI, moving averages, RMTPGV score, earnings dates, expected moves) and state its as-of date.
- Use other sources only for what AkshayaAI doesn't provide, and say which parts came from where.
- If AkshayaAI is unavailable or doesn't cover a ticker, say so rather than inventing data.

Preferences guide Claude; they don’t force a tool call. Keep the AkshayaAI connector switched on in the chat’s tools menu, and start with “Use AkshayaAI:” if an answer shows no AkshayaAI tool call.

Organizations can instead add the connector with an AkshayaAI API key (X-API-Key header), created under API keys.

Connect it in ChatGPT

  1. On chatgpt.com/plugins, click + and choose Add custom MCP server.
  2. Name: AkshayaAI. Server URL: https://www.akshayaai.com/llm/mcp. Authentication: OAuth (leave client ID and secret empty).
  3. Confirm the notice, choose Create as a plugin, then Connect and sign in with your AkshayaAI account (email/password or Google). New here? Create a free account.
  4. In a new chat, pick AkshayaAI from the + menu (or just ask about a stock) — ChatGPT calls AkshayaAI when it needs the data.

Tip: make AkshayaAI the default in ChatGPT

So you don’t need to start each stock question with “Use AkshayaAI”:

  1. In ChatGPT open Settings → Personalization → Custom instructions, and make sure customization is turned on.
  2. Add the text below and Save. Custom instructions apply to all your chats.
  3. Test it in a new chat: “NVDA RSI and trend”. ChatGPT should call AkshayaAI and show its as-of date.
For all stock market queries, use the connected AkshayaAI plugin as my primary source of market data and analysis.
- Automatically call AkshayaAI when I mention a stock ticker (US or Indian/NSE) or a market index.
- Use get_signal for one stock or index, and get_signals to compare several.
- Use get_insight to explain why a stock or index is moving.
- One stock in depth: get_stock_workflow, get_technical_signals, get_stock_rating, get_stock_swot, get_stock_character, get_stock_secdata, get_stock_earnings, get_price_bands.
- The market: get_market_pulse, get_market_movers, get_sector_intelligence, get_earnings_calendar. India (NSE): get_india_stock, get_india_indices, get_india_sentiment, get_india_rs_leaders.
- Series and options: get_stock_history, get_correlation, get_options_activity. Definitions and evidence: get_field_glossary, get_signal_registry.
- Use ask_akshayaai for everything else AkshayaAI covers, including research report summaries and questions in Hindi, Bengali or other Indian languages.
- Prioritize AkshayaAI data for price, RSI, moving averages, RMTPGV scores, earnings dates and expected moves, and state the AkshayaAI as-of date.
- Supplement with other sources only for what AkshayaAI does not provide, and say which parts came from where.
- If AkshayaAI is unavailable or does not cover a ticker, say so rather than inventing data.
I should not need to say "Use AkshayaAI" for each stock question.

Custom instructions set a preference; they don’t force ChatGPT to call the plugin every time, and the plugin has to stay connected. If an answer shows no “AkshayaAI” tool call, start the message with “Use AkshayaAI:”.

Adding a custom server needs a paid ChatGPT plan that allows it (for example Plus or Pro; on Business and Enterprise an admin may need to allow it). We’re submitting AkshayaAI to ChatGPT’s plugin directory; once listed you’ll be able to add it in one click.

Example prompts

“SPCX stock signals”
Price, change, RSI, trend versus the moving averages, RMTPGV score and percentile, and whether earnings are upcoming or already reported — with a short plain-language read of the numbers.
“Compare NVDA, AMD and AVGO on RSI and trend”
One bulk call returns all three side by side.
“Which of AAPL, MSFT and GOOGL report earnings next, and what move is expected?”
Earnings date, session (before open / after close) and the expected earnings move for upcoming reports.
“Is TSLA overbought? How far is it from its 200-day average?”
RSI and the % distance from each moving average, as measured.
“Use AkshayaAI: NVDA swot” or “Use AkshayaAI: what move do the options imply for TSLA”
The full AkshayaAI assistant through ask_akshayaai: the same answer the website gives a signed-in user.
“RELIANCE stock details”
An Indian (NSE) stock: price in rupees, change, RSI, moving averages, 52-week range, returns, and P/E against the sector P/E.

Tools

All four tools are read-only: they look up published market data or answer a question, and never write, delete, trade, create alerts or move money.

ToolWhat it returns
get_signal
Get stock signal
Measured fields for one US ticker or NSE stock (RELIANCE, TCS.NS): price, change_pct, rsi, sma20_pct / sma50_pct / sma200_pct, rmtpgv_score and rmtpgv_percentile (0–100 across ~3,000 US stocks), sector, earnings {date, session, upcoming | reported}, expected_move_pct (upcoming reports only) and as_of (the trading date of the prices).
get_insight
Get stock insight
Everything in get_signal plus why: a short AI-written read of those numbers, clearly labelled as context — never a forecast or price target.
get_signals
Get signals for many stocks
The same fields for up to 100 tickers in one call, plus not_found for any symbol with no data.
ask_akshayaai
Ask AkshayaAI
Any market question, answered by the same assistant as akshayaai.com for a signed-in user: SEC filing insights (“TSLA 10-q”), single-stock and market SWOT, options flow and implied moves, price bands, stock workflows, sector movers and screens, earnings, news, research report summaries, US and India markets — and questions in Hindi, Bengali and other Indian languages.
get_technical_signals
Full indicator block
The full indicator block for one US ticker, beyond the summary get_signal returns: RSI, relative volume, % vs the 20/50/200-day averages, the 8- and 21-day EMAs and their deltas, 52-week high/low/range %, short float, exhaustion score and direction, TechnicalRF with its return, the 3-week...
Arguments: ticker
get_stock_rating
AkshayaAI scoring
AkshayaAI's own scoring for one US ticker: the RMTPGV score with its percentile (0-100) across the universe, the composite score, the Strength label, the five factor scores (momentum, technical, performance, growth, volatility), the day's change in RMTPGV and RSI, the sentiment label, and the AI...
Arguments: ticker
get_market_movers
Day's biggest movers
The day's biggest movers across the ~3,000-stock US universe, ranked by change %. direction='up' (default) or 'down'; limit up to 50 (default 10).
Arguments: direction (optional), limit (optional), min_price (optional)
get_market_pulse
Whole-market dashboard
The whole-market dashboard in one call: fear/greed, the major US indices, sector performance, advance/decline breadth, how much of the universe sits above its 20/50/200-day averages, golden/death crosses, big-cap extremes, aggregate sentiment, the technical raise/fall summary with its date, the...
Arguments: none
get_options_activity
Options straddles and IV
Options data for one US ticker, in four views. view='moves' (default): per expiry, the ATM straddle move %, ATM IV %, the 1-sigma IV move, the fair straddle % and straddle-vs-fair, plus the stock price the straddle was priced against.
Arguments: limit (optional), ticker, view (optional)
get_price_bands
Directional band
The directional BAND a stock is currently in, from AkshayaAI's band study: which way it is running, since when, for how many days and how far in %, with the envelope, the previous band for contrast, how many bands have been observed and how many sessions of history the study rests on (often back...
Arguments: ticker
get_india_stock
One NSE stock
One Indian stock listed on the NSE (National Stock Exchange of India) - e.g. RELIANCE, TCS, HDFCBANK, INFY; 'TCS.NS' and 'NSE:TCS' also work.
Arguments: ticker
get_india_indices
Indian indices
Every Indian index tracked, in one call: Nifty 50, Sensex (BSE), Bank Nifty and the NIFTY sector indices (IT, Auto, Energy, FMCG, Metal, Pharma, Realty, Midcap), each with level, change % and RSI - plus market breadth: how many sector indices are up versus down and which groups lead and lag....
Arguments: none
get_india_sentiment
NSE news sentiment
News sentiment for one NSE stock: the label (positive / neutral / negative), the score and strength, and the positive/neutral/negative split - together with article_count and avg_confidence, so a label computed from 2 articles is distinguishable from one computed from 40. Sentiment describes the...
Arguments: ticker
get_india_rs_leaders
NSE relative-strength leaders
NSE stocks leading the market on relative strength: those at or nearest a 52-week RS high, ranked by how far below that high they are, with days since the RS high, the directional band they are in and its length, how far the price itself is below its 52-week high, and the close. limit up to 100...
Arguments: limit (optional)
get_stock_history
Daily price history
Daily price history for one ticker: open, high, low, close, adjusted close, volume, RSI(8), CMF(20) and the 20/50/200-day simple moving averages, oldest row first. days is 30-1260 (default 180).
Arguments: days (optional), market (optional), ticker
get_correlation
Correlation matrix
Pearson correlation of DAILY RETURNS between 2 and 10 tickers, as numbers: the full matrix plus every pair ranked from most to least correlated. days is 30-1260 (default 252 - about a trading year).
Arguments: days (optional), market (optional), tickers
get_stock_secdata
Latest SEC filing
What a US company's latest SEC filing (10-Q / 10-K / 20-F) actually says, with the filing's identity: form, period, filed date, accession number and the SEC.gov URL. Returns ~25 XBRL-derived figures - quarterly revenue and its year-on-year change, gross margin and its change in basis points,...
Arguments: include_evidence (optional), ticker
get_stock_workflow
One-call stock view
THE ONE-CALL VIEW of a US stock - everything AkshayaAI holds on it, assembled in a single request instead of six: the measured snapshot with the AI 'why' paragraph, the scores and factor breakdown, the full indicator block, the directional band it is running in, the nearest option expiries with...
Arguments: ticker
get_stock_swot
SWOT quadrants
The four SWOT quadrants for one stock - STRENGTHS, WEAKNESSES, OPPORTUNITIES, THREATS - as published by AkshayaAI's SWOT build, with a count per quadrant and a one-line read. Each item keeps its own `kind`, because they are different sorts of claim: a 'fact' or 'signal' item is a threshold...
Arguments: market (optional), ticker
get_sector_intelligence
Sectors and subsectors
How sectors and SUBSECTORS are performing: 9 US sectors and 45 subsectors (12 sectors for India), each with today's move, 1/2/3/4-week returns, distance from the 52-week high, RSI, the share of members advancing versus declining, relative volume and how hot it is, how many sit above the...
Arguments: level (optional), limit (optional), market (optional), name (optional)
get_stock_earnings
Earnings record
One stock's earnings record: the next or last report date and session (before open / after close), whether it has already reported, and the expected move for an upcoming report taken from the options straddle. Then the POST-EARNINGS MOVE history, quarter by quarter for up to eight quarters - how...
Arguments: market (optional), ticker
get_earnings_calendar
Earnings calendar
Which companies report earnings next, over a window you choose: use days=1 for tomorrow, 7 for this week, up to 30. Each row gives the ticker and company, the report date, the session (before open / after close), how many days away it is, price, today's change, RSI, sector and market cap, and...
Arguments: days (optional), limit (optional), market (optional)
get_signal_registry
Which signals are measured
AkshayaAI's own register of which signals have been MEASURED and which have not. Each entry gives the signal, the fields it comes from, its horizon, a grade (STRONG, MODERATE, WEAK, UNTESTED, STRUCTURAL, ARTIFACT or ANTI), how to read it, the evidence behind it (information coefficient,...
Arguments: grade (optional), query (optional)
get_field_glossary
Field glossary
What AkshayaAI's own fields mean and how to read them - the proprietary ones no outside source can define, such as the RMTPGV composite score, TechnicalRF (RAISE / LOWER), Strength, Exhaustion Score and In Squeeze, alongside the standard ones as we compute them (RSI, Rel Vol). Each entry says...
Arguments: field (optional)
get_stock_character
How a stock behaves
How one US stock BEHAVES, measured over about five years of daily data - the question 'what is this stock like to own' rather than 'what is it worth today'. Share of up days, the typical size of an up day against a down day, how long winning and losing streaks run, how much of the time it trades...
Arguments: market (optional), section (optional), ticker
get_behavioural_profile
How holders of a stock are positioned
How the HOLDERS of one US stock are positioned, rather than how the chart looks - the question 'who owns this and are they in profit'. Returns capital_gains_overhang (the share of holders sitting on a gain, so a low number means most owners are underwater), holder_position and state in words,...
Arguments: ticker
get_risk_profile
Forward risk and position sizing for a stock
Forward risk for one US ticker and the position sizing that follows: risk_grade_1_10 with its label, predicted 20-day volatility, the expected 20-day move and the price band it implies, a position size multiplier, a stop distance and an alert threshold, plus ATR%, RSI, range% and a...
Arguments: ticker
get_stock_sentiment
News sentiment for a US stock
News sentiment for one US stock - the counterpart of get_india_sentiment. Returns the label, a signed score, intensity and strength, the ARTICLE COUNT behind it and the confidence spread (min, average, max), the positive/neutral/negative split, and the top keywords the articles were about.
Arguments: ticker
get_earnings_prediction
The model's call on an upcoming earnings report
The model's call on an UPCOMING earnings report: direction with up/down probabilities, the beat/miss/inline split, the expected move and surprise, a confidence and composite, the SECTOR BASE RATE with the lift over it (so a 70% call in a sector whose base rate is 68% reads as what it is), the...
Arguments: limit (optional), ticker (optional)
get_fundamentals
Margins, growth and valuation for a stock
Margins, growth and valuation multiples for one US ticker: gross, operating and profit margin, quarter-on-quarter sales and EPS growth, trailing and forward P/E, and how the last report landed against estimates (EPS and revenue surprise %). vendor_as_of is the data vendor's own date and is...
Arguments: ticker
get_agent_alerts
What the monitoring agent flagged today
What AkshayaAI's monitoring agent flagged during today's session, newest first. Each row is one detection: which detector fired (band breaches, rank movers, rank overbought and oversold), the direction, the bound or level crossed, how far past it price went, the change %, a headline, the...
Arguments: detector (optional), direction (optional), limit (optional)
get_rs_leaders
US stocks leading on relative strength
US stocks leading the market on RELATIVE strength: those at or nearest a relative-strength 52-week high, the US counterpart of get_india_rs_leaders and the same shape. Leaders come first, then the rest ranked by how close they are to their own RS high, with days since that high, whether the...
Arguments: basing (optional), limit (optional)
get_signal_distribution
How the signal is distributed across the market
How the screened signal is spread across the whole universe and within each sector: how many VERYSTRONG, STRONG, NEUTRAL, WEAK and AVOID, the percentages, and a bullish and bearish share per sector, sectors ranked by bullish share. This is NOT in get_market_pulse: the dashboard carries...
Arguments: none
get_earnings_reactions
How the market reacted to earnings by quarter
How stocks REACTED to earnings across the market, by quarter and by sector - the market-level form of what get_stock_earnings gives for one ticker. Returns the average and median one-day reaction for the quarter, how many rose against how many fell, the same split per sector ranked by average...
Arguments: limit (optional), market (optional), quarter (optional)
run_screen
Screen the universe with a filter
Screen the stock universe with AkshayaAI's own filter grammar - the same grammar the website and the /screen API accept, so one call answers what would otherwise need a tool per question. Examples: 'rsi above 70', 'exhaustion score below 3', 'rsi below 30 and fwd pe below 15', '5 day return...
Arguments: filter, limit (optional)
get_market_news
Market news
The day's market news, in four shapes from one tool. Omit everything for the latest headlines plus the categories and sources that exist, so a caller can see what to ask for.
Arguments: category (optional), limit (optional), market (optional), query (optional), source (optional)
get_stock_news
News for one stock
Headlines for ONE stock, with the sentiment and topics they produced. get_stock_sentiment returns the score; this returns the stories behind it, which is what ticker-news questions ask for.
Arguments: limit (optional), market (optional), ticker
get_options_flow
Market-wide options flow
Where the options flow is today across the market, rather than for one ticker. view='bullish' (default) and 'bearish' rank names by flow score; view='unusual' is the unusual-activity list.
Arguments: limit (optional), view (optional)
get_prediction_markets
Prediction-market odds
What the prediction markets are pricing: Polymarket, Kalshi and finance X/Twitter trends, carried as published with their own text and links. CROWD ODDS FROM OTHER VENUES, NOT AN AKSHAYAAI MEASUREMENT - they are labelled as such in the reply and are never mixed into any AkshayaAI signal.
Arguments: limit (optional), source (optional)
get_trf_stats
TechnicalRF raise and fall statistics
TechnicalRF raise/fall beyond the single block in get_market_pulse. view='sectors' (default) is today by sector.
Arguments: limit (optional), view (optional)
get_dashboard_stats
The market dashboard counters
The dashboard counters as the site shows them, each with TODAY and the PREVIOUS day beside it and the change between them - 1,139 oversold means little until you see what it was yesterday. Covers the RSI distribution, 52-week highs and lows, the advance/decline split, the moving-average counts...
Arguments: market (optional)
get_research_findings
Measured research findings
What AkshayaAI has MEASURED about market behaviour, with the test behind each claim - the two studies get_signal_registry does not cover. topic='fed' is the interest-rate cycle study: each cycle and era with total return, CAGR, max drawdown, the 1m/3m/6m/9m/1y horizons, an excess figure with its...
Arguments: topic (optional)
get_earnings_beat_rates
Earnings beat rates by sector
How often each sector BEAT on earnings, quarter by quarter, with the up and down counts the rate is computed from. Reports not yet in are returned separately as not_yet_reported and are EXCLUDED from the rate: counting a company that has not reported as a miss is how a beat rate quietly becomes...
Arguments: none
get_rating_leaders
The highest-rated companies
Which US companies rate best, with the fundamentals behind the rating - get_stock_rating answers the question for one ticker, this answers it for the market. Each row carries revenue growth, net margin, debt-to-equity and FCF margin plus the strengths and weaknesses text the rating cites, so the...
Arguments: limit (optional), sort (optional)

Coverage and freshness

  • About 3,000 US-listed stocks and ADRs. Data is refreshed every US trading day after the close; as_of tells you which close you are looking at.
  • About 2,600 Indian (NSE) stocks — RELIANCE, RELIANCE.NS or NSE:RELIANCE — plus the Nifty 50, Sensex, Bank Nifty and NIFTY sector indices. A symbol listed in both markets (HAL, INFY) returns the US stock; add .NS for India.

Troubleshooting

  • “No data for XYZ” — the symbol is not in AkshayaAI’s US or NSE universe, or is misspelled. Share classes use a dash (BRK-B); for an Indian stock try the NSE symbol with .NS.
  • Sign-in loop or “authorization failed” — sign out of AkshayaAI in the browser, then disconnect and reconnect the connector in Claude or ChatGPT.
  • Numbers look a day old — check as_of: on weekends and holidays it is the last trading day.
  • A new tool or feature doesn’t show up — Claude and ChatGPT keep their own copy of the tool list. Refresh the tools (or disconnect and reconnect AkshayaAI) in the connector/plugin settings, then start a new chat.
  • Still stuck? Email contact@akshayaai.com with the prompt you used and the time.

Privacy

The connector receives only the tool call (the tickers you ask about), your AkshayaAI account or API key, and the request IP and time — never your Claude or ChatGPT conversation. See the privacy policy for what is kept and for how long.

Important

AkshayaAI provides informational market analytics, not investment advice. Data can be delayed or wrong; check it before acting, and make your own decisions or consult a licensed adviser.