create_forecast
ChatGPTSubmit a new energy-asset forecast simulation and return its run id. This creates a real run that counts against the account's forecast quota. simulation_config must satisfy the region's current input_json_schema exactly. Optional properties left out take the defaults the schema states. The run takes several minutes; results are fetched later by the returned id.
geocode_address
ChatGPTLook up site coordinates for an address or postcode inside a forecast region. Returns up to three matches inside the region, each with the matched address, latitude, longitude and precision: exact (the building), street, or area (a postcode or locality centre).
get_article_by_id
ChatGPTFetch full article content and chart data by ID. Use after article search or listing identifies a match via [source_id:X]. Includes structured chart data (title, type, axes, and raw data values) when the article contains charts. Use this data to answer questions about article visualisations — do not guess or describe charts you haven't fetched.
get_artifact
ChatGPTFetch a single artifact the authenticated user can open, including its full contents. Returns a short summary of the artifact, its file contents, and any chart data the body references for you to read, plus a structured reference that generates a link for the user to open. Or an error string.
get_asset_capacity
ChatGPTTotal rated power (MW), energy capacity (MWh) and asset count of assets over time, optionally broken down. Each row is the total on the first day of a period in [date_from, date_to). An asset counts on a date when its operational state that day is in operational_state.
get_benchmark_battery_rankings
ChatGPTRank benchmark batteries by deterministic revenue-per-MW over a preset window or a custom date range. Use when the user asks which batteries performed best or worst in the benchmark. Returns ordered rows plus the total ranked-population count for offset paging.
get_benchmark_values
ChatGPTGet time series value facts (e.g. revenue, revenue per MW, net power) for one or more benchmarks (asset, virtual asset, TB or index) over a given interval, optionally aggregated to a granularity. Accepts benchmarks published by Modo as well as ones created by you or within your company — pass the user's own asset name here just as you would a Modo one. Revenue facts: revenue is the raw total for the requested period (a one-month range returns one month of revenue). Use revenue_per_mw_year by default unless a user asks for a different unit. Workflow: find the benchmark with search_benchmarks or search_assets, then fetch values here by name — names are matched fuzzily, and the response notes any that didn't match exactly. Request all needed facts in one call — facts accepts several. Each search result's details lists the facts a benchmark supports. Links: when present, a "Links:" block after the summary gives each benchmark's page on Modo. Use for citations. Comparisons: to compare benchmarks — e.g. an asset against its region's Modo index (like ME BESS GB), or several assets against each other — pass all their names in names to get one merged table with a benchmark column, rather than fetching them separately and merging yourself. All benchmarks in one call must be in the same market region and support every requested fact (revenue facts work for all benchmark types). Markets: markets filters which markets are included (omit for all); breakdown controls the shape independently. By default (no breakdown) the selected markets are summed into a single row whose market column reads total. Pass breakdown="market" for one row per market plus that same summed total row (pinned last) — that is also how you discover which markets a benchmark trades, and lets you show breakdown and total together without summing per-market values yourself. Per-market responses end with a "Market glossary:" block describing each returned market, when descriptions are available. The market column holds human-readable market names ready to show the user; markets accepts those same names (or ids) matched fuzzily. Rows are per interval_start_local (bucket start in the region's timezone) as well when timeseries_granularity is set. Virtual asset results: the results facts (e.g. cashflow) come from a virtual asset's simulation, split into units such as the whole site, the grid_connection or one asset like assets.battery. Pass one unit per call. Values are in base units, not normalised. Rows are per dimension (e.g. a market, or the asset a flow comes from, or empty) and are never summed, so markets and breakdown don't apply. System benchmarks: ME BESS CAISO is a system benchmark. It shows fleet-wide performance, built from system-wide operational data, not individual asset results. It has no asset-level data, so an asset compared with it is not like-for-like. An asset may return no values for CAISO. Missing results: a benchmark that fails or returns no rows is named in the response text, and the values for the other benchmarks are still returned. Data lag: ERCOT revenue data is published with a ~60-day lag. For windows inside that lag, missing or partial data means "not published yet" — say so rather than reporting it as zero or absent revenue. Before offering or requesting analysis of ERCOT BESS revenue, benchmark performance, fleet performance, or per-asset performance, compare the requested period with today's date. Do not offer or request that analysis for a period ending within the previous 60 days because the results are not published yet.
get_bess_capacity
ChatGPTGet BESS capacity data (installed MW/MWh) from Modo Energy. Use when the user asks about battery storage capacity, installed MW/MWh, buildout, or growth. Returns power (MW) and energy (MWh) capacity. GB includes ALL assets (BMU and non-BMU); other regions exclude inactive assets. GB figures are month-start only — the next month's 1st IS the end-of-prior-period figure (end-2025 = 2026-01-01, not 2025-12-01).
get_forecast_custom_files
ChatGPTList custom file upload types or a company's uploaded files of one type. Called without file_type, this returns the types accepted by the forecasts service. With a type, it lists the caller's company files, newest first. The filename is the value to use in simulation_config; the upload ID and download URL are not config values. Uploading files is not supported here.
get_forecast_doc_page
ChatGPTRetrieve a whole forecast documentation page's content. Use when the user asks about a page named in a list_forecast_docs_pages row, or for any question a forecast documentation page answers.
get_forecast_nodes
ChatGPTGet node choices for a config field that takes a grid node. Lists the forecast node list. Search by name or code. Copy a selected option's value, not its label, into the applicable schema field: this may be a grid connection or a contract settlement point. For a node the user names, ask them to choose between ambiguous matches; never pick the first fuzzy match.
get_forecast_power_prices
ChatGPTGet Modo Energy's forecast power prices at a node as a time series. Returns the mean price of each period, with the period start on the forecast's clock. A list of scenarios, versions or nodes returns one dataset with a column that tells the series apart. One call returns fewer than 10,000 rows; when the range holds more, the response gives the time to continue from.
get_forecast_run_result
ChatGPTFetch result data for up to 6 forecast runs, or their status if not ready. While any run is still processing this returns run statuses instead of data; results become available once all requested runs reach status ready_for_review. Runs with status failed never produce results. When every run is ready, returns one table. With a single column, the table is a date column plus one column per run. With multiple columns (one run only), it is date plus one column per metric. Which metrics exist depends on each run's assets; units and currency vary by region and asset type. If a requested column does not exist for a run, the error lists that run's available columns.
get_forecast_runs
ChatGPTSearch existing energy-asset forecast runs (newest first). Use this to find runs when the user refers to past runs by name, date, region, scenario, model version, status or creator, or by how a run was configured. Status meanings: ready_for_review = results available; processing = not finished yet; failed = terminal, results will never exist. All filters combine with AND; list filters match any of their values. Results are every run in the caller's company library, not only the caller's own. Each run line shows its asset mix (battery/solar/wind/load), and the first 6 ready runs also list result_columns — the exact column labels their results offer (these vary by asset mix, region, and model version).
get_forecast_schema
ChatGPTGet the config schema for an energy-asset forecast. Forecasts model grid-scale batteries (BESS), solar, wind and on-site loads, alone or combined behind one grid connection. Called without market_region, this lists the released regions with their identifiers; one of those identifiers is what market_region takes. The response holds schema_version, the scenarios currently open for submission (exact values, including capitalisation), the region's valid start_date and end_date bounds, and input_json_schema, which defines the exact simulation_config object a forecast in that region accepts. It names each field that takes a zone or node, including nested asset contracts, and the tool that lists its choices. Only resolve fields relevant to the requested config. A property's description says what the input means and, where it applies, which other inputs change or replace its effect.
get_forecast_world_view_data
ChatGPTGet World View figures: Modo Energy's long-term outlook for a power market as yearly measures, such as demand, generation and capacity by technology, storage, price levels and spreads. Returns one row for each period and geography, with a column for each filter that has more than one value in the rows. A list of scenarios or versions returns one dataset with a column that tells the releases apart.
get_forecast_zones
ChatGPTGet zone choices for a config field that get_forecast_schema names. Copy market_region and layer from get_forecast_schema. Supply both site coordinates, or omit both to list choices for manual selection. The response includes value/label options and, when the site falls in a zone, a suggested one. Submit the selected value into the field. Re-fetch if coordinates change.
get_fx_rates
ChatGPTGet FX exchange rate estimates from the ECB data API. Call whenever currency conversion is needed. Every rate is a monthly average, never intraday or daily. For the reverse direction, request a new rate with the currency codes swapped; never invert a returned rate or reuse it in the opposite direction.
get_gb_solar_generation
ChatGPTGet GB solar PV generation data from Sheffield Solar via Modo Energy API. Use when the user asks about solar generation, solar output, or PV capacity in Great Britain. Granularity is auto-selected from the window length (half-hourly for ≤2 days, coarsening to monthly beyond a year). Data is past outturn only; there are no forecasts. Ask the user to clarify pes_region before calling if their location is ambiguous across regions. Example: "London total last week" → aggregation_type="sum" (MWh); "London average" → "mean" (MW).
get_help_docs_by_ids
ChatGPTRetrieve full help docs by their IDs. Use after list_help_docs to read the docs it lists. search_help_docs already returns full docs.
get_podcast_by_id
ChatGPTRead a podcast episode's full transcript. Use after retrieve_podcasts_list identifies the episode. If several episodes match, ask which one the user wants before loading a transcript.
get_tables
ChatGPTBrowse, search or inspect the database catalogue. Use summaries for data-availability questions. For retrieval, request schema detail directly, then run_sql. Use table_name to inspect a known table.
get_world_view_table_schema
ChatGPTDescribe one World View data table for a region. The response holds the metrics the table can return, with their units and descriptions, the filters it accepts, with the valid values for each read from the data, and the period range it covers. Geography is an ordinary filter — a geo_id at a geo_grain — so the data of the whole fundamentals market region is available, not only the geography behind one forecast region.
list_articles
ChatGPTList Modo Energy research articles by author, region, category and publication date. Use for publication lists, including buildout and revenue reports. These filters are not a free-text/keyword search. If the user's request is narrower than the available filters (e.g. a sub-market or topic within a region), narrow the returned list yourself by examining article titles. Results are capped at limit; when more articles match than the cap, the output reports the total — narrow the date window and re-query to retrieve the rest.
list_artifacts
ChatGPTList artifacts visible to the authenticated user, defaulting to artifacts they own. Returns each artifact's id, title, kind, file type, creation date and sharing; read one with get_artifact.
list_energy_academy_lessons
ChatGPTList the Energy Academy lesson catalogue. Returns lesson titles, creation dates, lesson references, and available course names and levels.
list_forecast_docs_pages
ChatGPTList all forecast docs pages (navigation breadcrumb, version, forecast_doc_id, docs_url). Use when the user asks where forecast documentation lives, which regions/countries the forecast covers, or wants the page index / table of contents. For forecast methodology, model, or input questions, prefer search_forecast_docs.
list_forecast_market_data
ChatGPTList Modo Energy's market forecast data: its forecast of each power market, published as releases. Called without market_region, this lists the regions that have data. For a region, the response holds the releases (each scenario with its versions, newest first), the power price series (its nodes, market types, granularities, time range and unit) and the World View tables, each with its filters and measures.
list_help_docs
ChatGPTList all available Modo help center docs. Use when the user wants to browse what help docs exist.
list_methodology_sections
ChatGPTList all section breadcrumbs for a benchmarking methodology page. Use to browse the document's structure or table of contents and select a section to retrieve. - AUS NEM Methodology (slug: australia) [url:https://modoenergy.com/benchmarks/methodology/asset/australia] - CAISO Methodology (slug: caiso) [url:https://modoenergy.com/benchmarks/methodology/system/caiso] - ERCOT Methodology (slug: ercot) [url:https://modoenergy.com/benchmarks/methodology/asset/ercot] - France Methodology (slug: france) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/france] - GB Methodology (slug: gb) [url:https://modoenergy.com/benchmarks/methodology/asset/gb] - Germany Methodology (slug: germany) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/germany] - Italy Methodology (slug: index-italy) [url:https://modoenergy.com/benchmarks/methodology/index/italy] - AUS WEM Methodology (slug: index-wem) [url:https://modoenergy.com/benchmarks/methodology/index/wem] - ISO-NE Methodology (slug: isone) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/isone] - Italy Methodology (slug: italy) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/italy] - MISO Methodology (slug: miso) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/miso] - NYISO Methodology (slug: nyiso) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/nyiso] - PJM Methodology (slug: pjm) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/pjm] - Poland Methodology (slug: poland) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/poland] - Spain Methodology (slug: spain) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/spain] - SPP Methodology (slug: spp) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/spp] - TBs Methodology (slug: tbs) [url:https://modoenergy.com/benchmarks/methodology/spread/tbs] - CAISO Methodology (slug: virtual-asset-caiso) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/caiso] - ERCOT Methodology (slug: virtual-asset-ercot) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/ercot] - Netherlands Methodology (slug: virtual-asset-netherlands) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/netherlands]
list_world_view_tables
ChatGPTExplore World View data — Modo Energy's long-term view of how each power market evolves. The horizon and the period grain differ by region and by table. Called without market_region, this lists the regions that have World View data; one of those identifiers is what market_region takes. For a region, the response holds the data tables available, the scenarios with their fundamentals versions, and how the region's geography is represented.
public_api_v1_get-endpoint
ChatGPTRead a Data API v1 or Forecasts API endpoint's HTTP request and response contract to write or debug the user's API integration. Returns documentation, not live data or a submitted forecast. Request examples must follow the documented types and omission defaults. When the contract specifies granularity=base for native intervals, include it explicitly; omission can aggregate the whole interval.
public_api_v1_search-endpoints
ChatGPTSearch Data API v1 or Forecasts API endpoint docs for raw Nord Pool, N2EX, EPEX, Elexon, NESO, Sheffield Solar PV Live, ERCOT and CAISO market feeds; Modo BESS asset databases, owners, leaderboards and per-asset revenue, also by index, for GB, ERCOT and the NEM; revenue and capacity of Modo and custom indices outside WEM; and forecast runs, input-file uploads and monthly or annual run results. Use it when the user wants this data in their own code or pipeline, or as a bulk export, download or CSV file. Returns matching endpoint paths and methods to read with get-endpoint. A keyword search miss does not prove absence. Use pattern '.*' to inspect this version's published endpoint catalogue before making a coverage claim.
public_api_v2_get-endpoint
ChatGPTRead a Data API v2 endpoint's HTTP request and response contract to write or debug the user's API integration. Returns documentation, not live data or a submitted forecast. Request examples must follow the documented types and omission defaults. When the contract specifies granularity=base for native intervals, include it explicitly; omission can aggregate the whole interval.
public_api_v2_search-endpoints
ChatGPTSearch Data API v2 endpoint docs for per-asset revenue, power, energy, throughput and cycles time series; revenue, capacity, members and per-asset metrics of Modo and custom indices in Italy and WEM; revenue time series of the user's virtual assets; the asset register (owners, capacity, grid locations and external IDs); and ENTSO-E day-ahead, EPEX day-ahead and intraday, and regelleistung.net aFRR and FCR prices. Use it when the user wants this data in their own code or pipeline, or as a bulk export, download or CSV file. Returns matching endpoint paths and methods to read with get-endpoint. A keyword search miss does not prove absence. Use pattern '.*' to inspect this version's published endpoint catalogue before making a coverage claim.
query_world_view_data
ChatGPTGet World View data for one region and table as a dataset. A filter value that is not in the data gives 0 rows; a scenario that does not resolve is an error. The dataset holds one row per period, and per dimension value where the table has dimensions. Unfiltered dimensions come back as columns so rows stay distinguishable; dimensions pinned by a single-value filter are omitted. A list of scenarios, geographies or filter values returns one combined dataset with a distinguishing column.
retrieve_energy_academy_lessons
ChatGPTRetrieve Energy Academy lesson transcripts containing beginner-friendly explanations of energy concepts. Returns matching lesson titles, creation dates, lesson references, and full transcripts. For lesson requests, recommend relevant matches even without an exact title match; do not substitute articles.
retrieve_methodology_section
ChatGPTRetrieve a full benchmarking methodology section, including all child subsections. Use when the user wants a section's full content or a search_benchmarking_methodology result is partial (e.g. [section chunk 1/3]) and more context is needed. - AUS NEM Methodology (slug: australia) [url:https://modoenergy.com/benchmarks/methodology/asset/australia] - CAISO Methodology (slug: caiso) [url:https://modoenergy.com/benchmarks/methodology/system/caiso] - ERCOT Methodology (slug: ercot) [url:https://modoenergy.com/benchmarks/methodology/asset/ercot] - France Methodology (slug: france) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/france] - GB Methodology (slug: gb) [url:https://modoenergy.com/benchmarks/methodology/asset/gb] - Germany Methodology (slug: germany) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/germany] - Italy Methodology (slug: index-italy) [url:https://modoenergy.com/benchmarks/methodology/index/italy] - AUS WEM Methodology (slug: index-wem) [url:https://modoenergy.com/benchmarks/methodology/index/wem] - ISO-NE Methodology (slug: isone) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/isone] - Italy Methodology (slug: italy) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/italy] - MISO Methodology (slug: miso) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/miso] - NYISO Methodology (slug: nyiso) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/nyiso] - PJM Methodology (slug: pjm) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/pjm] - Poland Methodology (slug: poland) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/poland] - Spain Methodology (slug: spain) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/spain] - SPP Methodology (slug: spp) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/spp] - TBs Methodology (slug: tbs) [url:https://modoenergy.com/benchmarks/methodology/spread/tbs] - CAISO Methodology (slug: virtual-asset-caiso) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/caiso] - ERCOT Methodology (slug: virtual-asset-ercot) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/ercot] - Netherlands Methodology (slug: virtual-asset-netherlands) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/netherlands]
retrieve_podcasts_list
ChatGPTSearch podcast episodes from Modo Energy Academy. Use for podcast, episode or audio requests. Read a selected episode's full transcript with get_podcast_by_id.
run_sql
ChatGPTRun a read-only Trino query and return the rows as CSV. Get schemas and source guidance with get_tables first. Prefer published benchmark tools for benchmark values. Preserve source coverage caveats. SQL - No SELECT . Quote column names. Only reference columns visible in the DDL. - Use current_date for "today". - Consider result set size. Aggregate (GROUP BY + AVG/SUM/COUNT/MIN/MAX) to summarise large datasets; raw rows only when explicitly requested. - For questions asking how high, low, or how much a metric moved, query its minimum, maximum, or range over the source's native intervals rather than inferring extrema from hourly or daily averages. - LIMIT is for top-N requests, never to shrink rowcount. - For window functions, compute over the full dataset then filter in an outer query — filtering before the window drops data. - Sort datetime columns ascending unless asked otherwise. - Period-keyed lookup tables: use the period containing the analysis dates; join multiple if the window crosses a boundary. - Times are market-local unless the user names a zone. Return each TIMESTAMP column as `at_timezone("<col>", '<zone>')`, never raw; for daily or longer buckets, convert before truncating and return a `DATE`. DATE columns are calendar dates: use them directly. Market zones: `'America/Chicago'` (ERCOT/SPP), `'America/Los_Angeles'` (CAISO), `'America/New_York'` (ISONE/MISO/NYISO/PJM), `'Australia/Brisbane'` (NEM), `'Europe/Berlin'` (GERMANY), `'Europe/London'` (GB), `'Europe/Madrid'` (SPAIN). - If the user asks for asset/technology-specific metrics and you only find market-wide tables, do NOT substitute — report that the breakdown is unavailable. Trino is NOT Postgres - `expr::TYPE` → `CAST(expr AS TYPE)`. - `ILIKE` → `LOWER(col) LIKE LOWER(pattern)`. - `INTERVAL '10' HOURS` → `INTERVAL '10' HOUR` (singular). - `EXTRACT(EPOCH FROM (a - b))` → `date_diff('second', b, a)`. - DECIMAL/DECIMAL division silently rounds (`5.0/60.0` returns `0.1`). Use scientific notation (`5.0E0/60.0E0`) or `CAST(... AS DOUBLE)`. - `sequence()` rejects zoned timestamps and returns at most 10,000 values. Build a time grid as `date_add('<unit>', <step> "n", <zoned start>) over UNNEST(sequence(0, <count>)) AS t("n")`.
search_articles
ChatGPTSearch Modo Energy's published articles by topic, including research, market reports and published directories of market participants. Returns relevance-ranked snippets. Use list_articles for a complete publication list.
search_assets
ChatGPTExplore visible assets with filters (region, state, technology, power, capacity, duration, commissioning date, company roles) and sorting. Each result is a flat, point-in-time snapshot of the asset as of as_of (today if omitted), mirroring the /v2/assets/{id} fields: id, name, market_region, technology, operational_state, rated_power_mw, energy_capacity_mwh, duration_hours, coordinates, commissioning_date, grid_locations, external_identifiers (e.g. BMUID, DUID), address_components, asset_groups, and technology_details (technology-specific attributes). When present, url is the asset's page on Modo; use for citations. map_url opens the asset on the Modo asset map. End your answer with a map link and invite the user to explore it: if the answer is about one or two assets, link each one's map_url; if it is about more, link the general asset map instead. also carries supported_facts — the fact names you can pass to get_benchmark_values for this asset, and id, which an index's assets filter takes (see get_benchmark_schema). The response echoes the effective as_of, so every snapshot's meaning is explicit; to compare across time, call once per date and diff. total_rated_power_mw and total_energy_capacity_mwh sum all total_count matches, not only the returned page, so a capacity total needs no paging. companies maps each role held on the asset as of as_of (e.g. owner, operator, optimizer, developer) to the names of every company holding it. Roles with no recorded holder are omitted.
search_benchmarking_methodology
ChatGPTSearch Modo Energy benchmarking methodology documentation. Use for battery revenue benchmark calculations, segmentation, definitions, regional performance metrics, the Benchmark Statement, and the Change and Cessation policy. Returns matching chunks with page slugs, section breadcrumbs and page URLs; partial chunks can be expanded with retrieve_methodology_section. - AUS NEM Methodology (slug: australia) [url:https://modoenergy.com/benchmarks/methodology/asset/australia] - CAISO Methodology (slug: caiso) [url:https://modoenergy.com/benchmarks/methodology/system/caiso] - ERCOT Methodology (slug: ercot) [url:https://modoenergy.com/benchmarks/methodology/asset/ercot] - France Methodology (slug: france) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/france] - GB Methodology (slug: gb) [url:https://modoenergy.com/benchmarks/methodology/asset/gb] - Germany Methodology (slug: germany) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/germany] - Italy Methodology (slug: index-italy) [url:https://modoenergy.com/benchmarks/methodology/index/italy] - AUS WEM Methodology (slug: index-wem) [url:https://modoenergy.com/benchmarks/methodology/index/wem] - ISO-NE Methodology (slug: isone) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/isone] - Italy Methodology (slug: italy) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/italy] - MISO Methodology (slug: miso) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/miso] - NYISO Methodology (slug: nyiso) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/nyiso] - PJM Methodology (slug: pjm) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/pjm] - Poland Methodology (slug: poland) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/poland] - Spain Methodology (slug: spain) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/spain] - SPP Methodology (slug: spp) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/spp] - TBs Methodology (slug: tbs) [url:https://modoenergy.com/benchmarks/methodology/spread/tbs] - CAISO Methodology (slug: virtual-asset-caiso) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/caiso] - ERCOT Methodology (slug: virtual-asset-ercot) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/ercot] - Netherlands Methodology (slug: virtual-asset-netherlands) [url:https://modoenergy.com/benchmarks/methodology/virtual-asset/netherlands]
search_benchmarks
ChatGPTFind benchmarks by fuzzy name and description. This searches benchmarks published by Modo, benchmarks created within your company, and benchmarks you created yourself. Each result is annotated with: - audience — modo (published by Modo, available to everyone), company (created within your company), or personal (created by you, visible only to you); - authored_by_you — whether you created this benchmark; - url — when present, the benchmark's page on Modo. Use for citations. When the user refers to their own asset or benchmark ("my PJM asset"), pass authored_by_you=true. When they mean Modo's published data, pass audience=["modo"]. The benchmark types are asset, virtual_asset, index, and tb; a handful of rows (e.g. ME BESS CAISO) have no subtype and report type: "benchmark". All are explorable here — this is the polymorphic discovery entrypoint for benchmarks data, so use it when you don't yet know which type a name like "ME BESS GB" refers to. Each result also carries a details block (dimensions and the names of the facts it supports), so you can call get_benchmark_values directly with a result's exact name (plus region if the name is ambiguous) — no separate lookup step. A query that exactly matches a benchmark name returns just that benchmark (not fuzzy neighbours), so passing a known name resolves it in one small response. Matching is lexical, not semantic: use words or fragments likely to appear in the benchmark name or description. Narrow with region rather than adding geography to the query. For multi-region comparisons, pass all the regions in one call. If a plausible name returns nothing, try market-code and duration-spelling variants before giving up.
search_forecast_docs
ChatGPTSearch forecast documentation for methodology, model, and input questions. Use as the first step when the user asks any forecast methodology question, what inputs or assumptions a forecast uses, how a specific region's forecast model works, or what the forecast can be run for — asset types and regions — including when the user mentions the Terminal. Note: the eastern-interconnection region covers US East markets — PJM, MISO, NYISO, and ISO-NE.
search_glossary
ChatGPTLook up an energy domain term or abbreviation.
search_help_docs
ChatGPTSearch Modo Energy help documentation for product-operation instructions. Use for Terminal navigation, screens and controls, data access and export steps, API or MCP connection setup, and account or preference settings. Includes how to open the forecast library or use a download button. Returns the full text of matching help docs.
skill_analyse-forecast-run
ChatGPTExplain the revenues of a forecast run, or the difference between two runs, with the market view behind them. Use when a user asks why a run earns what it does, why one run is higher or lower than another, what a setting in a run's config does, or what market assumptions sit behind a run.
skill_create-forecast
ChatGPTCreate a new forecast run for a grid-scale site with a battery, solar, wind, an on-site load (such as a data centre), or a mix of these behind one grid connection. Use when a user asks to run, submit, create, set up or kick off a new forecast.
skill_forecast-market-fundamentals
ChatGPTExplain why a market's power prices and spreads sit where they do in a Modo forecast, from fuel and carbon costs, generation and storage build, demand and transmission. Use when someone asks why forecast prices or spreads rise or fall, what drives the outlook for a market, or what explains a price change between forecast releases.
skill_forecast-revenue-analysis
ChatGPTExplain a battery's forecast revenues from the run's outputs and the market view behind them: which streams earn, why they move across the horizon, and what a comparison between runs shows, and what the model does when a user questions whether a forecast's level or trend is right or sets it against another provider's figures or realised outcomes. Loaded by forecast-revenue-report; use on its own for a forecast revenue question that needs the market and methodology evidence, not only the run figures.
skill_forecast-revenue-report
ChatGPTCreate a report on a battery's forecast revenues from a Modo forecast run: the revenue stack across the horizon, what drives it, and how it compares with an earlier release, a different configuration or another site when asked. Use when someone asks for a forecast revenue report, a write-up of their forecast results, what has changed since the last release, or a comparison of forecast options as a document.
skill_get-forecast-market-outlook
ChatGPTGet Modo Energy's long-term outlook for a power market from World View, the yearly figures of its market forecast, covering electricity demand, generation and capacity by technology, storage buildout, wholesale price levels and spreads, fuel prices and system reliability, under scenarios such as Low, Central and High. Use when a user asks what a power market does in future years, or asks about the market assumptions behind a forecast. Not for prices at a time step shorter than a year, and not for the revenues of a forecast run.
skill_get-forecast-power-price-data
ChatGPTGet Modo Energy's forecast power prices at a node as a time series, from prices at the forecast's own time step for a few days to monthly and yearly means across the forecast horizon, for one or more scenarios, releases or nodes. Use when a user asks for forecast prices at a time step shorter than a year, for the shape or the level of prices in a period, or for the prices behind a forecast run. Not for yearly market statistics such as spreads, and not for prices in the past.
skill_get-forecast-world-view-data
ChatGPTGet Modo Energy's long-term power-market outlook for a region — the summarised inputs and outputs of the Modo Energy power price forecast, covering electricity demand, generation mix and capacity, storage buildout, wholesale price levels and spreads, and system reliability, typically out to ~2050/2060, under scenarios such as Low, Central and High. Use when a user asks what a power market does in future years, or asks about the market assumptions behind a forecast.
skill_search-forecast-runs
ChatGPTFind and read existing forecast runs and their revenue results. Use when a user asks what runs exist, what a run earned, or to compare runs.