Acubic is a quant AI agent for investing. It screens assets, builds portfolios under your constraints, and monitors them through a broker connection.
This listing identifies a third-party MCP server. MCPBundles is not affiliated with, sponsored by, or endorsed by the service or its operator unless stated on this page.
This server may ask you to sign in when your AI client connects.
Last probed Nov 3, 2026 · mcp.acubic.ai
24tools discovered
For AI agents — MCPBundles Directory API
For MCP client apps (human connect):
MCP connect URL: https://mcp.acubic.ai/mcp
Skill page: https://www.mcpbundles.com/skills/acubic-mcp-3e21acb83f
Explore capabilities (24)
read 16
write 8
Acubic
24
Search Acubic knowledge
read
acubic_search_knowledge
Searches the content Acubic publishes on acubic.ai (blog articles, guides, the methodology page, the FAQ and pricing information) and returns matching passages with their links. Covers how Acubic builds strategies, what its backtests assume, plan limits and how the eToro and Trading 212 connections work. It does not look up individual stocks or funds and does not read any user data. Up to 10 results per page; a snippet is at most 400 characters (1,200 with response_format: detailed). Content is
Search Acubic knowledge — Searches the content Acubic publishes on acubic.ai (blog articles, guides, the methodology page, the FAQ and pricing information) and returns matching passages with their links. Covers how Acubic builds strategies, what its backtests assume, plan limits and how the eToro and Trading 212 connections work. It does not look up individual stocks or funds and does not read any user data. Up to 10 results per page; a snippet is at most 400 characters (1,200 with response_format: detailed). Content is
List strategy templates — Lists the presets Acubic offers when building a strategy: one for each combination of experience level (beginner, intermediate) and risk level (the largest drawdown the person says they can accept), plus Custom (advanced): the full engine, open to everyone. For each it returns the universe rules it applies, the ranking and weighting choices available, which pair Acubic marks as recommended, and the arguments to pass to acubic_create_strategy. A template describes what a setting does; it is not a
Look up an instrument — Looks up a stock, ETF, crypto asset or currency pair that Acubic knows, by ticker symbol, ISIN or name. Returns its name, ISIN, asset class, exchange and currency, whether it is in Acubic's tradable universe, whether eToro and Trading 212 list it, and any equivalent product Acubic substitutes where a broker lacks the US-listed one (for example a UCITS ETF for a US ETF). It returns identity and availability only: no price, return or rating. Indices and commodity price series are data-only and are
Screen the tradable universe — Applies filter rules (for example market cap at least $10 billion, P/E at most 25) to the instruments Acubic can trade, meaning those both eToro and Trading 212 can hold, and returns how many pass overall and per asset class plus up to 25 example rows ordered by market cap or by one of the metrics. Rules combine with AND, at most 20. Optionally narrows by asset class, region and sector first. Values come from the latest data snapshot (as_of); if the data cannot answer, the call fails rather than
Explain a term or setting — Explains one investing term or one Acubic setting in plain language: a financial metric (P/E ratio, beta, maximum drawdown), a filter, a ranking method, a weighting method, an experience level, a risk level, a rebalance frequency or a backtest term (walk-forward, benchmark, transaction costs). Returns a short explanation, how Acubic uses it, its default, common caveats and links to the Acubic page that covers it. It describes what the term means and how the option behaves; it does not say what a
Get your MCP into directories
A working endpoint is step one. Directory coverage is the coordinated launch across ChatGPT, Claude, Cursor, the MCP Registry, and community indexes.
Acubic is a quant AI agent for investing. It screens assets, builds portfolios under your constraints, and monitors them through a broker connection.
How do I connect Acubic to my AI agent?
Use the MCP endpoint listed on this page in your MCP client configuration. One-click install pills support Claude, Cursor, VS Code, and other hosts. Copy the remote MCP URL if your client needs a manual entry.
How many tools does Acubic provide?
Operate this server?
Request removal if you do not want Acubic listed here, or claim the listing if you want to manage it on MCPBundles.
Operate Acubic?
This server appears in the MCPBundles directory. Verify you operate it to take over the listing — name, description, logo, contact email, and skill content. We email a 6-digit code to a maintainer address your server publishes in /.well-known/security.txt or /.well-known/mcpbundles.json. Free, takes about a minute.
MCPBundles publishes this directory for MCP discovery. Except where we host or operate an offering, third-party services run under their own terms. Product and company names on this page are used in a descriptive, identifying way (including under nominative fair use where applicable); they remain the property of their owners. Nothing here grants you rights in those marks, and nothing here is an offer to sell a third party's services. Terms · Request removal
List strategy-builder options — Lists every choice Acubic's strategy builder offers, with a plain-language description, the trade-offs and the default at each experience level: risk levels, asset classes, regions, sectors, universe filters (30 metrics in 6 families), ranking methods, weighting methods, rebalance frequencies, and the backtest settings (years, transaction costs). Each option says whether it is currently offered. It describes options; the choice is the person's. Optionally limited to one section. Read-only, no si
Get my Acubic account — Returns the signed-in person's plan, how much of each monthly or daily allowance is used and when it resets, the live-portfolio limit and capital ceiling of the plan, their language and experience level, and a plan comparison with an upgrade link when an upgrade would raise a limit. It does not return the email address, payment details, broker credentials or identifiers from Stripe or the brokers. Read-only; no quota cost.
List my strategies — Lists the signed-in person's saved strategies, newest first, each with its name, whether it is a draft, a paper-money strategy or a live portfolio, the number of holdings, the ranking and weighting method, whether it has a saved backtest and when it ran, and a link to open it on acubic.ai. mode archived lists the ones deleted over MCP that can still be restored. It contains no performance figures; use acubic_get_strategy or acubic_get_backtest for those. Up to 25 per page. Read-only; no quota co
Get a strategy — Returns one saved strategy of the signed-in person: holdings and weights, the settings it was built with (experience level, risk level, ranking, weighting, asset count, filters, locks), the data date it was generated on, the summary of its latest saved backtest if there is one, and whether a broker is connected, which broker and how often it rebalances. It also accepts the build_id of an acubic_create_strategy call that has finished (a build still running is followed with acubic_get_build_status
Get a backtest result — Returns the latest saved walk-forward backtest of one of the signed-in person's strategies, or the status of a run in progress: the assumptions it used (years, rebalance frequency, transaction costs, ranking and weighting), CAGR, volatility, Sharpe and Sortino ratios, maximum drawdown and total return for the strategy and for the S&P 500 (SPY) over the same window, calendar-year returns, the largest drawdown episodes, and the holdings at the last rebalance. With include (or detailed) it also ret
Check a backtest's progress — Checks a backtest started by acubic_run_backtest without opening a card. While it runs: the percent done, a message and poll_after_seconds (wait at least that long, 15 seconds or more, before asking again). Once it has finished: the headline figures only, and the run's settings. A run that failed or was cancelled is an error and nothing was charged. Use it to wait; acubic_get_backtest returns the full result, once. A card already on screen updates itself and calls this tool on its own. Figures a
Check a strategy build's progress — Follows the build_id of acubic_create_strategy while the draft is still being assembled: percent done, a message and poll_after_seconds (15 seconds or more between asks). When the build has ended it returns the id of the saved strategy, which acubic_get_strategy reads, once. A failed or cancelled build is an error that cost nothing. The strategy card on screen follows the build on its own with this tool. Read-only; no quota cost.
Get my portfolio positions — Returns the holdings, value and last rebalance outcome of the signed-in person's live or paper-money strategies from the most recent daily snapshot Acubic stored (as_of gives its time). It does not contact the broker. It shows positions as percentages of the book, the book's value in the account currency, a time-weighted return since activation where the stored flows allow one, the outcome of the last rebalance as counts (placed, failed, waiting), and the next scheduled rebalance. It never retur
Create a strategy — Builds and saves a new strategy of the signed-in person from the acubic.ai builder's choices: asset classes, regions, sectors, number of assets, position sizes, ranking, weighting, filter rules, allocation and correlation limits, and tickers to add, exclude or lock in. A preset (template_id) fixes the menu; without one every engine argument is open. Acubic screens its tradable universe, orders and selects names, computes weights and saves a draft in the person's account. It places no order and c
Update a strategy — Changes the name, the description or the universe filter rules of one of the signed-in person's strategies. Replacing filter rules sets the strategy's metric rules to the list given (the asset-class, region and sector choices stay as saved) and requires the strategy to have been built with filters; at least as many names as its select_n (else its holdings count) must pass the new rules. It does not recompute holdings or weights, so a backtest run afterwards judges the edited rules. It cannot cha
Run a backtest — Runs a walk-forward backtest of one saved strategy of the signed-in person and saves the result on the strategy; it replaces the previous saved backtest, as only the latest is kept. At each rebalance date the strategy is re-built using only data available at that date, then held until the next, with the stated transaction costs, and compared with the S&P 500 (SPY). Settings: 1 to 10 years (default 5), monthly, quarterly or annual rebalancing (default monthly; annual needs at least 3 years), tran
Delete a strategy — Deletes one of the signed-in person's draft or paper-money strategies by archiving it together with its saved backtest and paper account: it leaves the person's strategies at once and can be restored within 30 days, after that it is removed for good. A strategy that has a broker connection cannot be deleted here and is removed on acubic.ai. The call asks the person to confirm; a client that cannot ask gets confirmation_required with a confirm_token (valid 5 minutes) and passes it back after the
Rank candidates — Ranks the candidates of a design before anything is saved (Order & Select on acubic.ai): returns the ranked list, page by page, with each candidate's rank, score, per-factor values (percentile and raw value), sector, country and asset class, the pool counts, the outcome of tickers added to the pool, and the pick the site would make: the top names, or those an allocation limit, a correlation limit or a hand-picked list gives. A fresh ranking screens first and uses one portfolio generation (upgrad
Regenerate the suggested portfolio — Regenerates the suggested portfolio of one of the signed-in person's saved strategies with its saved settings, as the Regenerate button of the acubic.ai portfolio page does: a filtered strategy is re-screened first, otherwise its pool is ranked again with its saved ranking, the names are selected under its saved limits and the weights computed with its saved method (it never changes the ranking or the weighting: for another one, build a new strategy, see acubic_create_strategy). Returns the sugg
Get a link to go live on acubic.ai — Returns a signed link, valid for 15 minutes, to continue one of the signed-in person's strategies on acubic.ai, where connecting a broker and starting automation happen. This tool does not connect a broker, place an order or start anything. The link opens the strategy's live setup page after sign-in; if the person's plan does not include live portfolios, it shows the plan options first. The result also states whether the plan includes live portfolios and its limits. Acubic runs live portfolios t
Project a strategy (scenario simulation) — Runs the AI Scenario Simulation of acubic.ai for one of the signed-in person's saved strategies: a Monte Carlo of 500 paths over 1, 3, 5 or 10 years, from its walk-forward backtest (else its in-sample figures). Returns the bear case (5th percentile), median and bull case (95th) at the horizon, the percentile fan and the assumptions; acubic_explain "Monte Carlo" describes the model. Carries a disclosure. Paid plans only (Free: upgrade_required with the plan comparison). Read-only; no quota cost.
Get a strategy's historical performance — Returns the Historical tab of acubic.ai for one of the signed-in person's saved strategies: the saved weights bought and held over 1 month to 5 years (default 1 year), with total return, CAGR, volatility, Sharpe ratio and maximum drawdown beside the S&P 500 and the daily series, downsampled. Optional custom benchmark of up to 10 weighted instruments. Carries a disclosure. Read-only; no quota cost.
Cancel a running backtest — Stops one of the signed-in person's walk-forward backtests that is still running, as Cancel on acubic.ai does: the allowance it used is returned and the previously saved result stays. Needs the strategy_id and the run_id from acubic_run_backtest. A run that already finished is reported and left as it is, so a repeat changes nothing.
Restore a deleted strategy — Brings back a strategy of the signed-in person that was deleted with acubic_delete_strategy, as the restore link on acubic.ai does, with its holdings, settings and saved backtest. Only within 30 days of the delete; acubic_list_strategies with mode archived lists the ones that can be restored. Repeating it is safe. No quota cost.
MCPBundles probed 24 tools on the live server. The tool list on this page reflects what was discovered at the last refresh — connect your client to see the full set available to your session.
What authentication does Acubic require?
Acubic may require signing in to the provider before tools can run. Connect through MCPBundles or your MCP client and complete any provider login when prompted.
Don't want this listed?
Request removal from the MCPBundles directory. We stop publishing this listing on mcpbundles.com.
MCPBundles is an independent platform built on the open Model Context Protocol standard. MCPBundles is not affiliated with, sponsored by, or endorsed by Anthropic PBC.