Ask in plain English. EconChat pulls from the authoritative source, renders a publication-ready chart, and traces every number to its API call. No portals. No spreadsheets. No guessing where the data came from.
Three portals. Three downloads. Merge in Excel. Build a chart. Format it. Someone asks where the numbers came from. You check your Downloads folder.
Generic chatbots skip the work but fabricate the answer. They mix up vintages, don’t know which source is authoritative, and fill gaps instead of flagging them.
EconChat is different. The LLM writes the narrative. The code pulls the data and runs the math. Never the reverse.
Not features. Deliverables.
Deliverable 1
Ask a question. Get a publication-ready chart with source badges, provenance, and data vintage. Follow up in the same conversation — context carries forward.
10 chart types. Automatic source routing. Every number traced to its API call.
Deliverable 2
Select a country. A full macro brief generates with GDP, inflation, debt, trade, and development indicators. Export as PDF or Word in your MDB’s house style.
5 institutional styles: World Bank, IMF, AfDB, IsDB, generic. Ready for circulation.
Deliverable 3
DSA (LIC-DSF), Growth Diagnostics (HRV), Structural Transformation (McMillan-Rodrik). Same methodology as your Excel template. Data auto-fetched. Peer benchmarks included.
The LLM writes the narrative. The code does the math. Never the reverse.
Before any response reaches you, six validators run against the raw data. When something doesn’t check out, you see exactly what and why.
Every number matched against raw API response
Overlapping indicators compared across databases
Empty API responses flagged, never filled with guesses
WEO edition verified — “April 2025”, not just “IMF”
No mixing “$2.3B” with “% of GDP” in the same claim
Trend claims (up/down/flat) checked against chart data
All claims verified against source data
UNCTAD FDI data unavailable for this period — flagged, not fabricated
Macro dashboard with KPIs, time-series charts, and development indicators. Data from WDI + WEO with full provenance.
GDP Growth
5.2%
Inflation
9.8%
Fiscal Balance
−4.1%
of GDP
Current Account
−2.3%
of GDP
Public Debt
68.4%
of GDP
Reserves
3.2
months
“Reconcile IMF and World Bank GDP data for Egypt.” Fetches the same indicator from both sources, computes point-by-point discrepancies, and recommends which to use.
“How has the IMF’s GDP forecast for Nigeria changed across WEO editions?” Pulls every vintage, renders a revision chart. No more downloading old Excel files.
Toggle full transparency. See every tool call, every API argument, every raw response. Know exactly which endpoint and indicator code was queried — not just the AI’s summary.
Open 3 data portals. Download CSVs. Merge in Excel. Hope the columns match.
One question. IMF, World Bank, and OECD merged automatically.
20 minutes formatting a chart. Screenshot it for the brief.
Publication-ready charts. Export as PDF, PNG, CSV, or Word.
“Where did this number come from?” Dig through Downloads.
Every number links to its source, API call, and data vintage.
Copy numbers from a staff report PDF. Retype what you can’t select.
Paste the URL. Tables extracted and structured automatically.
Run analysis in a shared Excel template. Hope nobody broke the formulas.
Same frameworks. Computed deterministically. Peer benchmarks included.
“Compare Ghana and Kenya GDP growth 2015–2024, include debt-to-GDP.” No query syntax. No portal navigation. Just the question.
Each indicator routes to its authoritative API. GDP forecasts from the IMF. Employment from the ILO. Trade from Comtrade. You don’t pick the database — EconChat already knows which one to trust.
Does every number match the raw data? Do sources agree? Did any API return empty? Is the WEO vintage cited correctly? Are the units consistent? Does the chart match the narrative? All six pass before the response reaches you.
Every feature. Every data source. Full provenance on every number.