BeeL vs PineForge — offline PineScript v6 backtest MCP
A side-by-side look at the BeeL and PineForge — offline PineScript v6 backtest MCP MCP servers: transport, auth, tools, quality signals, and the config to install either one in Claude, Cursor, Windsurf or VS Code.
Verdict: choose BeeL for finance & fintech work as a local process; choose PineForge — offline PineScript v6 backtest MCP for finance & fintech as a local process. You can also run both side by side in one mcpServers block.
Which MCP server should you choose?
Choose BeeL when
- You need dedicated capabilities in the Finance & Fintech domain.
- You prefer a local stdio subprocess transport.
- You have the keys it needs:
BEEL_API_KEY. - You want tools like
beel_activate_company, beel_cancel_representation, beel_change_managed_access_level.
Choose PineForge — offline PineScript v6 backtest MCP when
- You need dedicated capabilities in the Finance & Fintech domain.
- You prefer a local stdio subprocess transport.
- You have the keys it needs:
PINEFORGE_IMAGE, PINEFORGE_DOCKER_TIMEOUT_MS. - You want tools like
transpile_pine, backtest_pine, backtest_pine_grid.
Feature & specification comparison
Rows where the two servers differ are marked.
| Spec | ||
|---|---|---|
| Category | Finance & Fintech | Finance & Fintech |
| Quality signal | 63/100 · Good | 64/100 · Good |
| Install command | Confirmed | Confirmed |
| Transport | Local subprocess (stdio) | Local subprocess (stdio) |
| Auth | Not declared | Not declared |
| Pricing | Not declared | Not declared |
| Required env vars | BEEL_API_KEY | PINEFORGE_IMAGEPINEFORGE_DOCKER_TIMEOUT_MS |
| Verified owner | Community listing | Community listing |
| GitHub stars | 2 | 7 |
| Directory activity | 1 views 0 installs 0 upvotes | 1 views 0 installs 0 upvotes |
| Tools | 131 | 12 |
| Full listing | View BeeL | View PineForge — offline PineScript v6 backtest MCP |
Tools & capabilities
BeeL tools 131
beel_activate_companySwitches an existing company on in the mode carried in the body. The mode is always explicit and never taken from the credential's environment, so a Test key can switch a NIF on in Live. ## Modes and billing - **`TEST`:** immediate and free. - **`PROD`:** immediate when the account already has a card on file or an enterprise contract, and the NIF is added to the existing subscription. With no card on file it answers `402 CHECKOUT_REQUIRED`, returning a `checkout_url` when `success_url` and `cancel_url` are supplied. It also requires being the billing subject of the account (`403 NOT_BILLING_OWNER` otherwise). ## Idempotency and pending switch-offs - **Repeating the call:** opens no second checkout and adds no second subscription item; it returns the existing activation with `already_active: true`. The same `Idempotency-Key` sent to this route and to the nested one it replaces is the same operation, so it is replayed and never charged twice. - **A pending switch-off is cancelled:** while it is pending the NIF is still on — it just carries an effective date — so switching it on again only removes that date, answers `scheduled_deactivation_cancelled: true`, and charges or credits nothing. Endpoint: POST /v1/companies/{company_id}/activationsbeel_cancel_representationCancels the active AEAT representation of a company. - **Effect:** until a new document is generated and signed, the company can no longer submit invoices to AEAT in production. Its activation and its ability to issue non-VeriFactu invoices are untouched. - **No active representation:** rejected with `400`. Cancelling is a state transition, not a delete-if-present. Endpoint: DELETE /v1/companies/{company_id}/representationbeel_change_managed_access_levelUpdates the `access_level` you keep over an account you provisioned. - **Raising it:** only possible while the account is unclaimed. Once its holder has taken ownership you may keep or lower your access, but only they can raise it. - **Billing:** the level never affects it — you pay for the account's subscription at any level. - **`OPERATE`:** issuing invoices on the holder's behalf additionally requires a signed fiscal representation from them. - **Entitlement:** requires `manage_accounts`. Endpoint: PATCH /v1/accounts/{account_id}/access-levelbeel_convert_proforma_to_invoiceConverts an accepted proforma of this company into a real invoice. The new invoice is created as a `STANDARD` draft linked back through `source_proforma_id`. - **What converts:** only proformas in status `ACTIVE`. One shown as `EXPIRED` is still `ACTIVE` underneath and converts too. - **The proforma:** preserved as the record of what the customer accepted — it keeps its `PRO-...` number and PDF and moves to the terminal status `CONVERTED`. - **`issue`:** with `true` the new invoice is numbered and issued in the same atomic call. If issuing fails nothing is created and the proforma stays `ACTIVE`. That includes `400 SERIES_NUMBER_COLLISION`: the number the series would assign is already used by another invoice of the same company, and the series needs review, so contact support. - **Errors:** `422 CONVERSION_REQUIRES_PROFORMA` when the document is not a proforma, `422 PROFORMA_NOT_CONVERTIBLE` when it is not `ACTIVE`, and `409 PROFORMA_ALREADY_CONVERTED` when it has already been converted — a second call never creates a second invoice. Endpoint: POST /v1/companies/{company_id}/invoices/{invoice_id}/convert-to-invoice ⚠️ Read before calling: - The status names, the proforma lifecycle, and which tool performs each operation. (resource: beel://guardrails/invoice-state-machine)beel_create_claim_tokenIssues a single-use `claim_token`, and the `claim_url` built from it, so the account's holder can set a password and take ownership. - **`email`:** send it when the account has no holder yet — the person is created by this call. Omit the body to re-issue the token for the holder the account already has. An `email` that differs from the existing holder's is rejected rather than replacing them. - **Lifetime:** tokens last 30 days, and only the last one issued is live. Issuing again invalidates the previous token, so the old link stops working the moment you ask for a new one. - **Not an invitation:** this hands the account itself over to its holder. To add a further person to an account that already has one, invite them with `POST /v1/accounts/{account_id}/invitations`. - **Entitlement:** requires `manage_accounts`. Endpoint: POST /v1/accounts/{account_id}/claim-tokensbeel_create_invitationCreates a single-use invitation for a person to join the account with the given `account_role`. - **`token`:** the acceptance secret, returned once and never readable again, so deliver it to the invitee. `invitation_url` is the ready-to-use link built from that same token. - **`grants`:** the companies a `MEMBER` starts with. Omit it, or send `[]`, to invite them with no company access yet; an explicit `null` is rejected with `422`. Grants are only valid for `MEMBER`, since `OWNER` and `ADMIN` reach every company implicitly. - **`account_role`:** `OWNER` cannot be invited. An account has exactly one owner, handed over only through `PUT /v1/accounts/{account_id}/owner`. - **`send_email`:** defaults to `false`, so BeeL sends no email and you deliver the token or `invitation_url` yourself. Set it to `true` to have the invitation emailed to `invited_email` as well. Endpoint: POST /v1/accounts/{account_id}/invitationsbeel_create_webhook_subscriptionRegisters an HTTPS endpoint to receive notifications for the event types listed in `events`. - **`secret`:** returned **only** in this response and never again. Store it before discarding the body; deliveries are signed with it and carry the signature in the `BeeL-Signature` header. - **`test_delivery`:** a one-off signed delivery sent to your URL as part of creating the subscription, so you learn whether your endpoint answers without a second call. It is best effort: the subscription exists and is active whatever it says, and the field is `null` when the test could not be run at all. - **`account_relationship`:** which accounts the subscription receives events from — `own` (the default), `managed`, or `all`. - **Limits:** an account holds at most **10 active subscriptions**; creating an eleventh is rejected with `400` `WEBHOOK_ACTIVE_SUBSCRIPTION_LIMIT_REACHED`. Registering the same URL twice creates two subscriptions, and the endpoint then receives each event twice. Endpoint: POST /v1/accounts/{account_id}/webhooksbeel_create_companyCreates a company under the account the request resolves to. The NIF is registered in the name of that account's holder, never in the name of the caller. - **`activate`:** unless it is `false`, the company is switched on in `aeat_environment` and its three default invoice series (ordinary, simplified, corrective) are seeded there. This endpoint never switches an existing company on: that is `POST /v1/companies/{company_id}/activations`. - **`numbering`:** decides the code, format, counter reset and starting number those series are born with. Only accepted when the request activates the company. - **Billing:** no charge is ever started here. Creating a production NIF requires being the billing subject of the account (`403` otherwise), and an account without billing is rejected with `402`; no checkout is opened in either case. - **Duplicates:** a NIF that already exists in the account is rejected with `409`, and the response carries the existing `error.details.company_id`. - **Addresses:** a Spanish postal code (`country_code` omitted or `ES`) must have 5 digits, in `address` and in `legal_representative.address`; otherwise `422 POSTAL_CODE_INVALID_ES` and nothing is created. Other countries' postal codes are free-form. Endpoint: POST /v1/accounts/{account_id}/companies ⚠️ Read before calling: - Which company an operation acts on, and how that is selected. (resource: beel://guardrails/multi-nif) - Why a name that does not match the census makes an invoice unsubmittable. (resource: beel://guardrails/nif-validation) - How a series formats numbers, and which series configurations are rejected. (resource: beel://guardrails/series-and-numbering)beel_create_corrective_invoiceIssues a corrective invoice that amends the invoice in the path. It is a new fiscal document with its own number, not an edit of the original. - **`rectification_type`:** `TOTAL` leaves the original `VOIDED` and rectifies what is still invoiced on it: every line of the original and of its live correctives (voided ones do not count), negated. It takes no `lines` — sending them fails with `422 RECTIFICATIVA_TOTAL_CON_LINEAS`. `PARTIAL` leaves the original `RECTIFIED` and requires the adjustment `lines`. - **Never more than was invoiced:** a `PARTIAL` may raise any amount, but may not take the taxable base of any rate (tax, rate and equivalence surcharge; `SUPLIDO` lines by their amount) below zero once the previous correctives are counted. That fails with `422 CORRECTIVE_EXCEEDS_INVOICED_AMOUNT`, and `error.details` (`CorrectiveInvoiceErrorDetails`) carries `tax_group` (for example `IVA 21%`) and `max_reduction`, how much of that rate is left to rectify. A `TOTAL` on an invoice that previous correctives already brought to zero fails with `422 CORRECTIVE_NOTHING_LEFT_TO_RECTIFY`. - **Not for the withholding alone:** a `PARTIAL` whose lines leave the taxable base of every rate unchanged and only change the withholding fails with `422 CORRECTIVE_WITHHOLDING_ONLY`. A withholding is not a cause for a corrective: void the invoice and issue a new one without it. - **The original's PDF:** unchanged by either type. The corrective has its own PDF; the original keeps the one that was delivered, and its new status is in `status`. - **Total of 0:** a corrective whose `total_to_pay` is 0 has nothing to refund or collect, so it is issued as `PAID`, with `payment_date` equal to `issue_date`. - **What can be rectified:** an ordinary or simplified invoice in `ISSUED`, `SENT`, `PAID`, `OVERDUE` or `RECTIFIED`. Rectifying a corrective fails with `422 CORRECTIVE_NOT_RECTIFIABLE` — to fix an erroneous corrective, issue another one against the original invoice. - **Repeat rectifications:** several `PARTIAL` correctives are allowed, but a `VOIDED` invoice is no longer rectifiable, so a second `TOTAL` against the same invoice fails with `422 INVOICE_NOT_CORRECTIBLE_IN_CURRENT_STATUS`. - **Correcting the recipient's data:** when the invoice recorded its recipient with a wrong name, tax ID or address, send the corrected `recipient` with `rectification_type` `PARTIAL`, `rectification_code` `R4` and no `lines`. The corrective carries the corrected recipient and does not change the amounts: its lines negate what is still invoiced and repeat it, so every rate nets to zero. A different person is not a data correction (`422 CORRECTIVE_RECIPIENT_IS_ANOTHER_PERSON`): correct the invoice in full and issue a new one to the right customer. - **Original rejected by the AEAT:** when VeriFactu rejected the original's record and it has not been resubmitted, the original is not in the AEAT's books: fix and resubmit it first. Until then the request fails with `422 CORRECTIVE_ORIGINAL_RECORD_REJECTED`. - **Deadline:** four years from when the tax accrued (the original's operation date) or, for a cause of article 80 of the VAT Act, from the `circumstance_date` you declare. Past it the request fails with `422 CORRECTIVE_OUT_OF_TIME`, with `deadline` and `counted_from` in `error.details` (`CorrectiveInvoiceErrorDetails`). - **What the reason code requires** (Ley 37/1992, art. 80): `R2` (insolvency) and `R3` (bad debt) need a recipient established in Spain, the Canary Islands, Ceuta or Melilla —an `R2` also accepts a recipient in another EU member state, for insolvency proceedings there— and fail otherwise with `422 CORRECTIVE_RECIPIENT_NOT_ESTABLISHED`. An `R3` needs at least six months since the original's operation date (`422 CORRECTIVE_BAD_DEBT_TOO_EARLY`, with `earliest_date`; one year when the previous year's turnover exceeded 6,010,121.04 €, which is the issuer's to apply), and on an operation with a base of 50 € or less it needs `recipient_is_business` (`422 CORRECTIVE_BAD_DEBT_BASE_TOO_LOW`). The other conditions of each code (claims, guarantees, related parties, filing with the AEAT) are the issuer's to meet. - **Fiscal inheritance on a `PARTIAL`:** a line that omits `irpf_rate` or `equivalence_surcharge_rate` takes it from the **original invoice** — the document being amended — and never from the company's current tax profile, so a profile that changed after the original was issued does not leak into the credit note. An explicit value always wins, `0` included. The surcharge inherits the *regime* (on/off), not the rate: the rate is re-derived from each corrective line's own VAT (21→5.2, 10→1.4, 5→0.62, 4→0.5), and an original outside the regime pins the line to `0`. `SUPLIDO` lines are out of it on both sides. When the original is not unambiguous BeeL does not pick for you: different IRPF rates per line fail with `422 CORRECTIVE_ORIGINAL_MIXED_IRPF`, and a surcharge applied on some lines but not others fails with `422 CORRECTIVE_ORIGINAL_MIXED_SURCHARGE`. Declare the figure on every line to get past either — both only fire when some line actually needs to inherit. - **`series_id`:** when omitted, the document is numbered in the company's default corrective series, never in the series of the original: corrective invoices go in a series of their own (RD 1619/2012, art. 6.1.a). If the company has none, it is created on first use (code `R`, or the next free one that cannot repeat another series' numbers). An explicit `series_id` must be a corrective series. - **Numbering conflict:** if the number the series would assign is already used by another invoice of the same company, in this series or in another one, the request fails with `400 SERIES_NUMBER_COLLISION` without issuing anything or consuming a number. The series needs review, so contact support. Endpoint: POST /v1/companies/{company_id}/invoices/{invoice_id}/corrective ⚠️ Read before calling: - Fiscal rules, domains corrective, void: beel_rules_list with domain, or resource beel://guardrails/<domain>. - How a line states its price, and which field combinations are rejected. (resource: beel://guardrails/invoice-lines) - The status names, the proforma lifecycle, and which tool performs each operation. (resource: beel://guardrails/invoice-state-machine) - How a series formats numbers, and which series configurations are rejected. (resource: beel://guardrails/series-and-numbering)beel_create_customerCreates a new customer under this company. - **`Idempotency-Key`:** it identifies the same operation on the deprecated flat route, so a retry that switches route replays instead of creating twice. Endpoint: POST /v1/companies/{company_id}/customers ⚠️ Read before calling: - Why a name that does not match the census makes an invoice unsubmittable. (resource: beel://guardrails/nif-validation)beel_create_customers_bulkCreates up to 500 customers of this company in a single call. - **Atomic:** if any customer fails validation the whole batch is rejected with `422` `BULK_VALIDATION_ERROR` and nothing is persisted. This is not a partial operation. - **`dry_run`:** with `dry_run=true` the batch is only validated — tax identifiers against the AEAT register, duplicates inside the batch and against the existing customers, field formats — nothing is written and the answer is `200`. With `dry_run=false`, the default, validation is followed by creation and the answer is `201`. - **Report:** both modes return the same per-record report, so a dry run and a real run are read the same way. Endpoint: POST /v1/companies/{company_id}/customers/bulk ⚠️ Read before calling: - Why a name that does not match the census makes an invoice unsubmittable. (resource: beel://guardrails/nif-validation)beel_create_invoiceCreates an invoice for this company. The issuer data comes from the company in the path, and the document is created as a draft unless you ask for it to be issued. - **Issuing:** `options.issue_directly` numbers and issues the invoice in the same call. Submission to the AEAT is asynchronous, so `verifactu.submission_status` comes back as `PENDING`: a 2xx means the invoice was accepted for submission, not that the AEAT has registered it. If the number the series would assign is already used by another invoice of the same company, in this series or in another one, it fails with `400 SERIES_NUMBER_COLLISION` without issuing anything or consuming a number: the series needs review, so contact support. - **Document type:** `type` chooses the document. A `PROFORMA` is non-fiscal — it is born `ACTIVE`, numbered `PRO-...` from its own non-fiscal series, and ignores `issue_directly`. - **Related:** to copy an existing invoice into a new draft, use `POST …/invoices/derivations`, which carries neither `type`, nor `recipient`, nor `lines`. Endpoint: POST /v1/companies/{company_id}/invoices ⚠️ Read before calling: - Fiscal rules, domains simplified, contents, taxes, surcharge: beel_rules_list with domain, or resource beel://guardrails/<domain>. - How a line states its price, and which field combinations are rejected. (resource: beel://guardrails/invoice-lines) - Why a name that does not match the census makes an invoice unsubmittable. (resource: beel://guardrails/nif-validation) - How to tell, before issuing, whether a NIF can issue, and what each blocker means. (resource: beel://guardrails/verifactu-gates) - How a series formats numbers, and which series configurations are rejected. (resource: beel://guardrails/series-and-numbering)- +119 more on the listing page
PineForge — offline PineScript v6 backtest MCP tools 12
transpile_pineTranspile PineScript v6 source to a C++ translation unit locally, using the pineforge-codegen transpiler bundled in the pineforge-release Docker image. No API key, no network — source never leaves the machine. Returns the generated C++ as text. Use backtest_pine if you also want to run the strategy.backtest_pineRun a real, deterministic backtest of a PineScript v6 strategy — prefer this over estimating its trades or P&L by reasoning, which is unreliable for Pine (series semantics, intrabar fills, and strategy.* order logic do not reproduce from approximation). Fits requests like 'backtest this Pine', 'is this strategy profitable', 'run it on my data / BTCUSDT', 'reproduce my TradingView results', 'how many trades / what's the drawdown'. Transpile a PineScript v6 strategy and run it against an OHLCV CSV via the pineforge-release Docker image on the user's local machine. Fully local — transpile + backtest run in-container; nothing leaves the box, no API key. Optional `inputs` overrides input.*() named values from the Pine source (keys = the second arg of input.*(...) calls, e.g. 'Fast Length'). Optional `overrides` overrides strategy(...) header fields (initial_capital, commission_value, default_qty_value, pyramiding, slippage, default_qty_type, commission_type, process_orders_on_close). Returns the parsed JSON report (summary, trades, applied_inputs, applied_overrides, applied_runtime, elapsed_seconds). The instrument matters: pass `symbol` (a Binance symbol) or `syminfo` (your own qty_step, mintick, ...); a CSV from fetch_binance_ohlcv carries its instrument and needs neither. The lot size is TradingView's own reading for the symbol, from a measured table shipped with this server (Binance's LOT_SIZE.stepSize only for a symbol TradingView does not list; TradingView's usual 0.001 for a listing newer than the table); the tick size and currencies come from Binance's public exchangeInfo (or from the sidecar next to a CSV fetched by fetch_binance_ohlcv). `syminfo` goes over all of it. What was applied is in applied_runtime.syminfo; when the lot grid is unknown, or its lot size is not a TradingView reading, the result has a `warnings` entry. If the report is too large to return inline it is written to report_path and a compact summary (with that path) is returned instead. Use backtest_pine_grid for sweeps.backtest_pine_gridUse when the user wants to optimize, sweep, tune, or compare PineScript parameter values (e.g. 'try fast length 8/12/19', 'find the best commission/qty settings') rather than test a single configuration — for one fixed configuration use backtest_pine. Run a parameter sweep: transpile the Pine source ONCE (locally, in-container), then compile (g++) and backtest that C++ against the OHLCV CSV once per combination in the cartesian product of `inputs` × `overrides` grids. Returns a ranked list of {inputs, overrides, summary, elapsed_seconds} entries sorted by `sort_by` descending, plus the top entry under `best`. Cap: max_combinations (default 64). Takes the same `symbol` / `market` / `syminfo` as backtest_pine and applies the instrument to every combination (the result's `instrument` shows it). The lot size is TradingView's own reading for the symbol, from a measured table shipped with this server (Binance's LOT_SIZE.stepSize only for a symbol TradingView does not list; TradingView's usual 0.001 for a listing newer than the table); the tick size and currencies come from Binance's public exchangeInfo (or from the sidecar next to a CSV fetched by fetch_binance_ohlcv). `syminfo` goes over all of it. Set concurrency > 1 to run backtests in parallel — each docker container has its own startup overhead, so 2-4 is usually plenty.check_tradingview_parityCheck how closely PineForge reproduces a TradingView backtest, trade by trade. Give the Pine v6 script and TradingView's own Strategy Tester export: the "List of trades" CSV, or the XLSX report as base64. PineForge runs the script on the same market and window and grades the two trade lists with the grader behind its recorded parity figures (pineforge-engine v1.2.0 scripts/verify_corpus.py, byte-identical in v1.3.0 and the frozen 1.4.0 candidate). The execution harness run_strategy.py is copied byte-for-byte from frozen engine candidate b3192bfc2f5a24bf4efd6d1d01e01e8fe619dfed; file identity is not a new native corpus parity result. Returns the tier (excellent, strong, moderate, weak, minimal), each check with its value and thresholds, matched and unmatched trade counts, the first mismatches side by side with hints, and a timezone check. Bars: pass ohlcv_csv or ohlcv_csv_path for any market; without them, BINANCE:<SYMBOL> (spot) and BINANCE:<SYMBOL>.P (USDT-M perpetual) bars are fetched from Binance's public API (at most 100,000 bars); any other symbol needs your bars. No quota and no history window beyond that. Scripts declaring use_bar_magnifier=true also fetch 1-minute bars through the last chart bar's close, counted in that limit, on charts the harness magnifies (coarser than 1 minute, at most 1 day) unless runtime.bar_magnifier is false. With your own bars, optionally pass magnifier_ohlcv_csv or magnifier_ohlcv_csv_path; a script that declares the magnifier but runs without one gets a warning that fills inside bars may differ from TradingView's. Settings the XLSX Properties sheet states are used; an explicit input that disagrees with one is an error. The run applies the instrument's lot size and tick size as backtest_pine does, for BINANCE:<SYMBOL> and BINANCE:<SYMBOL>.P. The lot size is TradingView's own reading for the symbol, from a measured table shipped with this server (Binance's LOT_SIZE.stepSize only for a symbol TradingView does not list; TradingView's usual 0.001 for a listing newer than the table); the tick size and currencies come from Binance's public exchangeInfo (or from the sidecar next to a CSV fetched by fetch_binance_ohlcv). `syminfo` gives your own values and goes over all of it. The result's `applied_instrument` shows what was applied, and `warnings` says when no lot size could be found or its lot size is not a TradingView reading. The run and the grading use the pineforge-release Docker image (no network inside the container).fetch_binance_ohlcvFetch OHLCV candles from Binance public API and write a backtest-ready CSV (header: timestamp,open,high,low,close,volume; timestamp = open time in UNIX ms UTC). Supports `spot` and `usdt_perp` (USDT-margined perpetual futures). Requests larger than 1000 bars are paginated automatically. Also records the symbol's instrument in <output_path>.instrument.json (lot size: TradingView's own reading from a measured table shipped with this server, Binance's LOT_SIZE.stepSize for a symbol TradingView does not list, TradingView's usual 0.001 for a listing newer than the table; tick size and currencies from Binance's public exchangeInfo), which backtest_pine and backtest_pine_grid pick up; the result's `instrument` shows what was recorded. The output path must live inside the MCP cwd unless PINEFORGE_ALLOW_ANYWHERE=1.binance_symbolsList/validate symbols available on the Binance public API for OHLCV fetching. Filters: `query` (substring of the symbol), `quote_asset` (e.g. 'USDT'), `base_asset` (e.g. 'BTC'), `status` (e.g. 'TRADING'), `contract_type` (futures only, e.g. 'PERPETUAL'). Results are cached 5 min in process. Free.list_engine_paramsReturns the full catalog of engine knobs accepted by backtest_pine / backtest_pine_grid in two groups: strategy_overrides (the 9 strategy(...) header fields the runtime reads via PINEFORGE_OVERRIDES — initial_capital, pyramiding, slippage, commission_value, commission_type, default_qty_value, default_qty_type, process_orders_on_close, close_entries_rule) and runtime_args (input_tf, script_tf, bar_magnifier, magnifier_samples, magnifier_dist — args to run_backtest_full, NOT part of the strategy() header). Each entry is {key, type, enum?, description}. Does not run the engine. Use this to discover what knobs the engine exposes before issuing a backtest.list_coverage_topicsSTART HERE before writing, porting, or backtesting a Pine v6 strategy on PineForge. PineForge implements a SUBSET of Pine v6, so checking coverage first avoids a strategy that compiles but silently misbehaves vs TradingView. Lists every coverage topic with a one-line status (supported / partial / unsupported / via_transpiler) and summary, plus the legend (note: via_transpiler still works end-to-end; unsupported means refused, not compiling, stopping the run where its value is read, or accepted with no effect) and the coverage version. Statuses describe what a backtest on THIS server can do. Cheap, free, local — no engine run, no I/O. Then drill in with get_coverage_topic for one area's per-feature lists, or check_pine_feature to look up a single identifier.get_coverage_topicReturns the full detail plus the exact supported[], partial[], via_transpiler[] and unsupported[] feature lists for ONE coverage topic id (ids from list_coverage_topics, e.g. 'ta', 'strategy_orders', 'request_security', 'drawing_plotting_alerts'). Use when you are about to work in a feature area and need to know precisely which functions there are implemented vs skipped — e.g. before using request.security, the ta.* library, or strategy risk knobs. Unknown ids return an error marker listing the valid ids. Local, free, no engine run.check_pine_featureAnswer "does PineForge support X?" for a specific Pine v6 identifier or namespace (e.g. 'ta.supertrend', 'alert', 'array.new', 'request.financial'). Use it (a) BEFORE relying on any function you are unsure about while writing a strategy, and (b) to DIAGNOSE a backtest that compiled but behaved wrong or empty — plots, tables and alerts (plot, bgcolor, table, alert) are accepted and produce NO effect, while line, box and label objects are data the strategy can read back. Resolves by exact feature match, then longest namespace prefix, then alias, returning {query, status, topic, note} where status is supported / partial / unsupported / via_transpiler / not_found (via_transpiler = works end-to-end; unsupported = refused, not compiling, stopping the run, or no effect) and the note quotes the catalog entry, which says what THIS server can and cannot do (e.g. request.security on another symbol). Local, free, no engine run.pull_engine_imageRun `docker pull` for the pineforge-release runtime image on the user's machine. Useful before the first backtest_pine call.check_engine_imageCheck whether the local pineforge-release Docker image is up to date with the registry. Compares per-platform manifest digests via `docker manifest inspect --verbose` (no image layers downloaded). Returns up_to_date + recommend_pull. With auto_pull=true, runs `docker pull` in the same call when the local image is stale or missing. Note: this is independent of the MCP server's own version (`@pineforge/backtest-mcp`); the MCP version and the engine image version evolve separately.
Ready-to-paste client configs
Paste either block (or both) into your client config, e.g. claude_desktop_config.json or ~/.cursor/mcp.json. We only show a config when the install command is confirmed.
BeeL
{
"mcpServers": {
"beel": {
"command": "npx",
"args": [
"-y",
"@beel_es/mcp"
],
"env": {
"BEEL_API_KEY": "YOUR_BEEL_API_KEY_HERE"
}
}
}
}
PineForge — offline PineScript v6 backtest MCP
{
"mcpServers": {
"pineforge-offline-pinescript-v6-backtest-mcp": {
"command": "npx",
"args": [
"-y",
"@pineforge/backtest-mcp"
],
"env": {
"PINEFORGE_IMAGE": "YOUR_PINEFORGE_IMAGE_HERE",
"PINEFORGE_DOCKER_TIMEOUT_MS": "YOUR_PINEFORGE_DOCKER_TIMEOUT_MS_HERE"
}
}
}
}
Frequently asked questions
BeeL is categorized under Finance & Fintech and uses a local stdio subprocess. PineForge — offline PineScript v6 backtest MCP belongs to Finance & Fintech and uses a local stdio subprocess. Pick BeeL for finance & fintech work and PineForge — offline PineScript v6 backtest MCP for finance & fintech.
Copy the mcpServers block from the config section on this page into your client's config (for example claude_desktop_config.json or ~/.cursor/mcp.json), then fully restart the client.
BeeL has no declared pricing — check the project's own site. PineForge — offline PineScript v6 backtest MCP has no declared pricing — check the project's own site. If a server needs environment variables such as API keys, define them under the env key in your client's MCP config.
Yes. MCP clients support multiple servers: add beel and pineforge-offline-pinescript-v6-backtest-mcp as separate keys in the same mcpServers object, and your assistant will call tools from either one.
On AllMCPs, BeeL has a quality signal of 63/100 (Good) with 1 views, 0 installs and 2 GitHub stars. PineForge — offline PineScript v6 backtest MCP has 64/100 (Good) with 1 views, 0 installs and 7 GitHub stars.