Live release feed
Sub-second macro releases for FX backtests
Point-in-time history
Official CPI, jobs, GDP, and central-bank events with point-in-time history.
$25/month 14-day free trial
Start Free Trial
Free prompt library 19 prompts MCP native

Prompt library for FX macro agents.

Copy repeatable workflows for FX macro analysis, event-risk planning, pair research, visual charts, and quant-style checks. Every prompt is designed to route agents through FXMacroData MCP instead of stale model memory.

Browse prompts Open JSON catalog Connect MCP

Free

Prompts are free to copy. Public MCP calls cover setup, USD public rows, FX rates, calendars, sessions, and catalog discovery.

MCP native

Several prompts map directly to MCP prompts/get, including macro briefings, pair analysis, and week-ahead plans.

Guardrailed

The library asks agents to separate data from interpretation, cite timestamps, label gaps, and avoid personalized financial advice.

Upgradeable

Non-USD indicators, COT, commodities, app visuals, task tools, and backtests may require OAuth or an API key.

Machine-readable

One library, multiple agent surfaces

Use the page for humans, JSON for apps, and plain text for LLM retrieval. The canonical MCP endpoint remains https://fxmacrodata.com/mcp.

JSON catalog /prompt-library.json Plain text /prompt-library.txt LLM prompt feed /llms-prompts.txt Well-known JSON /.well-known/fxmacrodata-prompts.json

Prompt catalog

Copy a workflow into your agent

Replace bracketed placeholders such as [BASE], [QUOTE], and [CURRENCY] before running the prompt.

Getting Started

Connect an agent, verify tool access, and make the model cite FXMacroData rows instead of guessing.

Free

MCP Connection Smoke Test

Verify that the host can see FXMacroData MCP tools before asking for analysis.

ping data_catalogue indicator_query release_calendar
View full prompt 
Use the FXMacroData MCP server at https://fxmacrodata.com/mcp. First call ping. Then call data_catalogue(currency="usd"). If inflation is available, call indicator_query(currency="usd", indicator="inflation", limit=3). Then call release_calendar(currency="usd") for the next few rows. Return a compact connection report with tool names called, whether data was returned, and any access limitations. Do not infer values that did not come back from FXMacroData.

Output sections

  • Connection
  • Visible Tools
  • Returned Data
  • Access Notes

Best for

  • ChatGPT connector setup
  • Claude/Cursor/Codex MCP validation
  • Support triage
Guardrail: Use this before pasting a long research request into a new agent host. Access: No paid key is needed for the smoke test.
Free

Agent System Instruction

Give a general AI agent a stable source hierarchy and safety contract for FX macro work.

data_catalogue indicator_query release_calendar forex market_sessions
View full prompt 
You are an FX macro research assistant using FXMacroData as the source of record. Use FXMacroData MCP tools before model memory whenever the user asks for live, recent, historical, scheduled, or structured macro and FX data. Call data_catalogue before uncertain indicator requests. Use indicator_query for macro rows, release_calendar for upcoming event risk, forex for spot history, market_sessions for trading hours, cot_data for CFTC positioning, and commodities for commodity series. State the currency, indicator, date range, and tool used. Cite announcement_datetime for macro figures. If a tool returns no data, say so plainly. Do not fabricate values. Provide neutral research context, not personalized investment advice, trade instructions, guaranteed outcomes, or position sizing.

Output sections

  • Source Hierarchy
  • Tool Routing
  • Citation Rules
  • Safety Boundary

Best for

  • Custom GPT instructions
  • Claude project instructions
  • Codex/Cursor workspace rules
Guardrail: Keeps the assistant focused on data-grounded analysis and away from personal advice. Access: Protected tools still require OAuth or an FXMacroData API key.
Free plus paid data_quality_audit

Data Quality Audit

Make the agent inspect freshness, source, coverage, and missing-data fields before summarizing.

data_catalogue indicator_query release_calendar
View full prompt 
Audit FXMacroData coverage for [CURRENCY] [INDICATOR]. Call data_catalogue(currency="[CURRENCY]") to verify the slug. Then call indicator_query(currency="[CURRENCY]", indicator="[INDICATOR]") and release_calendar(currency="[CURRENCY]", indicator="[INDICATOR]"). Inspect mcp_metadata, result.data_quality, latest_available_date, source fields, row count, and any pagination or no-data flags. Return an audit note that separates reliable facts, missing fields, stale data, and follow-up checks. Do not produce market interpretation until the audit section is complete.

Output sections

  • Coverage
  • Freshness
  • Source Metadata
  • Data Gaps
  • Follow-Up Checks

Best for

  • Before publishing charts
  • Before backtests
  • Before client notes
Guardrail: Prevents a fluent answer from hiding stale or incomplete data. Access: Non-USD and full-history checks may require paid access.

Briefings

Daily and weekly macro routines for analysts, traders, and research desks.

MCP prompt macro_briefing

Daily Macro Briefing

Create a repeatable one-page macro brief for a single currency.

data_catalogue indicator_query release_calendar
View full prompt 
Produce a one-page macro briefing for [CURRENCY]. Call data_catalogue first. Then retrieve policy_rate, inflation, unemployment, GDP, and the next scheduled releases with FXMacroData MCP tools. Use only returned values. Cite announcement_datetime for every figure. Keep interpretation neutral and label any missing series.

Output sections

  • Current Stance
  • Inflation Path
  • Labor Market
  • Growth
  • Next Releases

Best for

  • Morning desk note
  • Client meeting prep
  • Single-currency research thread
Guardrail: Briefing language should be market context, not a recommendation. Access: Use an API key for non-USD indicators and deeper history.
MCP prompt week_ahead_macro_plan

Week-Ahead Macro Plan

Build a structured week-ahead plan around a focus FX pair.

release_risk_score_task macro_research_pack_task market_sessions
View full prompt 
Build a week-ahead macro plan for [BASE]/[QUOTE]. Use FXMacroData release_risk_score_task for the pair, macro_research_pack_task for both currencies using policy_rate as the anchor, and market_sessions for timing. Rank events by relevance, map them to FX sessions, and produce neutral scenarios. Do not present the result as trade advice or a forecast guarantee.

Output sections

  • Highest-Priority Events
  • Session Timing
  • Scenarios
  • Data Gaps

Best for

  • Weekly planning
  • Persistent agent threads
  • Event-risk calendar reviews
Guardrail: Scenarios should be conditional and data-grounded. Access: Task tools and non-USD data typically require paid access.
Free

Session-Aware Market Plan

Map macro event timing to Sydney, Tokyo, London, and New York trading sessions.

market_sessions release_calendar forex
View full prompt 
Create a session-aware market plan for [BASE]/[QUOTE] today. Call market_sessions() first. Then call release_calendar for both currencies and forex(base="[BASE]", quote="[QUOTE]") for recent spot context. Summarize which sessions overlap with scheduled risk, which data rows were used, and what conditions would make the day unusually event-heavy. Keep the output as a planning memo, not a trade instruction.

Output sections

  • Open Sessions
  • Scheduled Releases
  • Spot Context
  • Planning Notes

Best for

  • Intraday desk prep
  • Timezone planning
  • Release-window awareness
Guardrail: Do not tell the user to trade a session; identify timing and context only. Access: Most session and FX spot checks are available without a key.

Pair Analysis

FX-pair prompts for spot context, rates, inflation, positioning, and scheduled risk.

MCP prompt pair_analysis

FX Pair Analysis

Analyze a pair through spot, technicals, rates, inflation, and positioning.

forex indicator_query cot_data
View full prompt 
Analyze [BASE]/[QUOTE] using FXMacroData. Call forex with SMA/RSI context, policy_rate for both currencies, inflation for both currencies, and cot_data where supported. Return a neutral macro view with the exact dates and values returned by the tools. If COT or non-USD data is unavailable, say so and continue with available public context.

Output sections

  • Spot Context
  • Rates
  • Inflation
  • Positioning
  • Neutral Macro Bias

Best for

  • Pair research
  • Custom GPT prompt cards
  • Analyst workbench threads
Guardrail: Use 'macro bias' rather than buy/sell instructions. Access: COT and non-USD indicator access may require paid access.
MCP prompt pair_risk_check

FX Pair Risk Check

Triage the biggest current and scheduled risks for a currency pair.

pair_intel_task release_risk_score_task macro_regime_classifier_task
View full prompt 
Run a risk check for [BASE]/[QUOTE]. Use pair_intel_task, release_risk_score_task, and macro_regime_classifier_task for each leg. Separate current setup from upcoming risk. Explain what data would change the view, and identify missing coverage. Do not provide personalized investment advice, position sizing, or execution instructions.

Output sections

  • Current Setup
  • Top Upcoming Risks
  • Regime Divergence
  • Invalidation Conditions

Best for

  • Risk review
  • Before an event-heavy week
  • Portfolio watchlists
Guardrail: Use conditional risk framing rather than directives. Access: Pair intelligence and regime tools are subscriber-oriented workflows.
Free plus paid

Carry And Rate Differential Check

Compare policy-rate paths and carry pressure across an FX pair.

indicator_query forex policy_rate_differential_visual_artifact
View full prompt 
Compare carry and policy-rate pressure for [BASE]/[QUOTE]. Use indicator_query for policy_rate on both currencies and forex for recent spot context. If the host supports MCP Apps, prefer policy_rate_differential_visual_artifact for the chart. Explain the current rate spread, recent direction of the spread, and whether spot has moved with or against the rate backdrop. Use dates and values from FXMacroData only.

Output sections

  • Current Spread
  • Recent Rate Path
  • Spot Context
  • Carry Caveats

Best for

  • Carry research
  • Rate-differential dashboards
  • Pair setup notes
Guardrail: Carry is one input, not a standalone trade recommendation. Access: Full multi-currency policy-rate history requires paid access.
Paid data cot_positioning_context

COT Positioning Context

Use CFTC positioning as a positioning-extreme overlay for supported currencies.

cot_data forex pair_intel_task
View full prompt 
Summarize COT positioning context for [CURRENCY] and relate it to [BASE]/[QUOTE]. Call cot_data(currency="[CURRENCY]") and forex(base="[BASE]", quote="[QUOTE]"). Report latest report date, net speculative position, recent changes, and whether positioning appears crowded or light relative to recent history if the response supports that comparison. Label unavailable COT fields clearly.

Output sections

  • Latest Report
  • Net Positioning
  • Recent Change
  • Pair Context
  • Caveats

Best for

  • Positioning notes
  • Risk review
  • Cross-checking consensus narratives
Guardrail: Do not treat COT as a timing signal by itself. Access: Requires an FXMacroData API key when COT is protected.

Event Risk

Central-bank, CPI, payrolls, GDP, and post-release workflows with point-in-time guardrails.

MCP prompt rate_decision_brief

Central-Bank Rate Decision Brief

Summarize policy-rate stance, inflation, labor context, and next central-bank meeting timing.

indicator_query release_calendar
View full prompt 
Produce a central-bank rate-decision brief for [CURRENCY]. Call indicator_query for policy_rate, inflation, and unemployment. Then call release_calendar(currency="[CURRENCY]", indicator="policy_rate") for the next scheduled meeting. Cite announcement_datetime for every figure and never invent unavailable data.

Output sections

  • Current Stance
  • Inflation Path
  • Labor Market
  • Forward Guidance
  • Next Meeting Date

Best for

  • Decision-day prep
  • Central-bank monitoring
  • Policy-rate watchlists
Guardrail: Keep the output to data-backed scenarios and context; do not claim to know the decision outcome. Access: Use an API key for non-USD indicators and deeper policy-rate history.
MCP prompt pre_central_bank_decision_brief

Pre-Central-Bank Decision Brief

Prepare a decision-preview memo before a rate announcement.

macro_regime_classifier_task indicator_query release_calendar macro_research_pack_task
View full prompt 
Prepare a pre-central-bank decision brief for [CURRENCY]. Use macro_regime_classifier_task, policy_rate, inflation, release_calendar for policy_rate, and macro_research_pack_task. Frame scenarios around data and timing. Do not claim to know the decision outcome; do not recommend a trade.

Output sections

  • Market Setup
  • Regime Classification
  • Decision Risk
  • Scenarios
  • Next Timestamp

Best for

  • FOMC/ECB/BoE/RBA prep
  • Client meeting notes
  • Agent memory threads
Guardrail: Decision previews must stay scenario-based. Access: Most decision-preview workflows need paid access for full context.
MCP prompt post_release_trade_summary

Post-Release Research Summary

Summarize what changed after a macro release without overstating causality.

event_impact_replay_task macro_research_pack_task pair_intel_task
View full prompt 
Write a post-release research summary for [CURRENCY] [INDICATOR] using [BASE]/[QUOTE] as market context. Use event_impact_replay_task, macro_research_pack_task, and pair_intel_task. State what printed, how it compared with recent context, what the pair did around the event if available, and what remains uncertain. Use announcement_datetime and avoid claiming a single-cause market move unless the returned data directly supports it.

Output sections

  • What Printed
  • Prior Context
  • Pair Context
  • Uncertainties
  • Next Watch

Best for

  • Post-CPI notes
  • Payrolls recap
  • Release-journal entries
Guardrail: Avoid causal overreach; label context as context. Access: Event replay and pair intelligence are protected workflows.
Free plus paid release_risk_ladder

Release Risk Ladder

Rank upcoming releases for a pair and tie them to trading-session timing.

release_calendar release_risk_score_task market_sessions
View full prompt 
Build a release-risk ladder for [BASE]/[QUOTE] over the next [N] scheduled events. Call release_calendar for both currencies. If available, call release_risk_score_task(base="[BASE]", quote="[QUOTE]"). Use market_sessions to map release timing to sessions. Return a ranked list with timestamp, currency, indicator, why it matters, and what data would reduce uncertainty.

Output sections

  • Ranked Events
  • Session Mapping
  • Why It Matters
  • Data Gaps

Best for

  • Calendar triage
  • Week-ahead watchlists
  • Agent-generated event calendars
Guardrail: Ranking event risk is not the same as predicting price direction. Access: Calendars work publicly; risk scoring is a protected workflow.
Paid workflow known_at_time_backtest_check

Known-At-Time Backtest Check

Prevent lookahead bias before an agent writes a strategy note or backtest.

known_at_time_task indicator_query
View full prompt 
Check what FXMacroData would have known about [CURRENCY] [INDICATOR] as of [AS_OF_DATE]. Use known_at_time_task(currency="[CURRENCY]", indicator="[INDICATOR]", as_of="[AS_OF_DATE]"). Compare the result to a normal indicator_query only after the known-at-time slice is shown. Explain which rows were market-visible by the as-of timestamp and which rows must be excluded from any backtest.

Output sections

  • As-Of Context
  • Visible Rows
  • Excluded Rows
  • Backtest Rule

Best for

  • No-lookahead audits
  • Backtest prompts
  • Strategy research reviews
Guardrail: Never use future-released rows in historical reasoning. Access: Known-at-time checks are subscriber-oriented tools.

Quant Workflows

Backtest, portfolio, commodity, visual, and data-quality prompts for reproducible agent work.

Paid workflow event_gated_backtest_brief

Event-Gated Backtest Brief

Ask an agent to test a simple FX rule with macro release windows instead of hand-waving.

fx_backtest_task release_calendar known_at_time_task
View full prompt 
Create an event-gated backtest brief for [BASE]/[QUOTE] from [START_DATE] to [END_DATE]. Use fx_backtest_task with strategy="carry_momentum", event_gated=true, event_window_days=1, and realistic transaction_cost_bps. Use release_calendar and known_at_time_task only where needed to explain timing. Report assumptions, sample size, metrics, drawdowns, and failure modes. Label all results as hypothetical research, not live performance or investment advice.

Output sections

  • Assumptions
  • Metrics
  • Event Gate Impact
  • Failure Modes
  • Research Caveats

Best for

  • Quant notebooks
  • Strategy ideation
  • Event-window testing
Guardrail: Hypothetical backtests are research aids, not performance promises. Access: Requires protected backtest and point-in-time workflows.
Paid workflow portfolio_event_risk_scan

Portfolio Event-Risk Scan

Scan a small FX book for scheduled macro concentration and stress exposure.

portfolio_risk_engine_task release_risk_score_task market_sessions
View full prompt 
Scan this FX position list for macro event risk using FXMacroData: [POSITIONS_JSON]. Use portfolio_risk_engine_task with a clear stress_shock_pct and horizon_events. If needed, use release_risk_score_task for the highest-risk pairs and market_sessions for timing. Return concentration, upcoming event exposure, stress assumptions, and monitoring priorities. Do not recommend changes to the user's actual portfolio.

Output sections

  • Positions Parsed
  • Concentration
  • Upcoming Event Exposure
  • Stress Assumptions
  • Monitoring Priorities

Best for

  • Desk risk reviews
  • Watchlists
  • Pre-event exposure checks
Guardrail: Describe risks; do not tell the user to add, reduce, or close positions. Access: Requires protected portfolio-risk tooling.
Paid data commodity_fx_context

Commodity FX Context

Connect commodity moves to commodity-sensitive FX pairs without using third-party runtime data.

commodities commodities_visual_artifact forex indicator_query
View full prompt 
Analyze commodity context for [BASE]/[QUOTE] using FXMacroData only. Call commodities(indicator="[COMMODITY]") and forex(base="[BASE]", quote="[QUOTE]"). If relevant, call indicator_query for terms_of_trade, commodity_price_index, inflation, or policy_rate where available. Explain the commodity path, pair context, and macro channels. Do not use external market-data providers.

Output sections

  • Commodity Path
  • Pair Context
  • Macro Channels
  • Data Limits

Best for

  • AUD/CAD/NOK context
  • Commodity shock notes
  • Cross-asset research
Guardrail: Treat commodity linkage as context, not deterministic causality. Access: Commodity series and many non-USD indicators require paid access.
MCP prompt visualize_macro_series

Visualize Macro Series

Force chart requests toward MCP Apps visual artifacts when the host can render them.

indicator_visual_artifact forex_visual_artifact commodities_visual_artifact cot_visual_artifact
View full prompt 
Visualize this FXMacroData request: [REQUEST]. If the host supports MCP Apps, prefer the relevant visual artifact tool for indicators, FX, commodities, or COT. Return the chart plus a short interpretation that names the tool, date range, latest value, and any stale or missing data flags. If the host cannot render the app, return a compact text fallback.

Output sections

  • Chart
  • Latest Value
  • Trend
  • Data Quality
  • Fallback

Best for

  • ChatGPT Apps
  • Claude/Cursor visual hosts
  • Dashboard-style research
Guardrail: A chart is a visual summary, not a trade signal. Access: Visual artifact tools are usually protected beyond public USD evaluation.

Safety boundary

Research prompts, not trading signals

These prompts are designed to make agents fetch official-source macro data, inspect timing and freshness, and produce neutral analysis. They should not be used to generate personalized investment advice, position sizing, guaranteed outcomes, or automated trading instructions.

Docs Connect FXMacroData MCP ChatGPT Open Custom GPT