OpenEcon Data¶
external · status: active · focus: analysis · discipline: economics · started: 2026
Project page: https://github.com/hanlulong/openecon-data
Licence: unclassified (licence file present; see repository)
Source: projects/landscape/openecon-data.yml
Positioning¶
An MCP server plus web app giving AI agents natural-language access to 330K+ economic indicators indexed across 11 providers (FRED, World Bank, IMF, Eurostat, UN Comtrade, BIS, Statistics Canada, OECD, ExchangeRate-API, CoinGecko, ChinaMacro). Sits in the data-acquisition layer of RISE, complementary to StatsPAI/stata-code (which analyze data the agent already has) rather than overlapping with them.
Distinctive contribution¶
The only economic-data-acquisition MCP tool in the catalog: an LLM-parser → semantic-provider-routing → fetch-with-fallback pipeline that resolves natural-language requests (including multi-language queries and automatic regional-group expansion like G7/BRICS/EU) into indicator series, with conversational follow-ups and multi-format export (CSV/JSON/Stata/Python code) carrying source attribution.
Evaluation scores¶
| Dimension | Score (0–3) | Note |
|---|---|---|
| Lifecycle coverage | 0 | Touches one stage only (data-acquisition); no analysis, drafting, or review functionality. |
| Autonomy level | 0 | Tool: every query is human- or agent-issued; conversational follow-ups persist context but do not amount to independent task execution. |
| Architectural transparency | 2 | Three-stage pipeline (LLM parser → semantic routing → fetch-with-fallback) documented and open-sourced (FastAPI/React/Redis stack); no published prompts for the LLM-parser stage. |
| Inputs supported | 1 | One input form (natural-language query) plus data-source access across 11 providers; no literature-corpus access. |
| Outputs / reproducibility | 1 | Persists structured data exports (CSV/JSON/Stata/Python) with source attribution and cached deterministic re-fetch; no paper/code/data-manifest bundle — it is a data tool, not a paper-producing pipeline. |
| Internal evaluation | 0 | Only latency metrics reported (~0.1s cached, several seconds first query); no evaluation of query-to-indicator mapping accuracy. |
| Openness | 1 | AGPL-3.0 (copyleft, source-sharing required for hosted service); free hosted demo gated at 20 queries before signup. |
| Maturity / traction | 2 | 81 stars, 891 commits, live production demo (data.openecon.ai/chat) — active beta with external users, single-maintainer team. |
| Cross-family policy | 0 | Not applicable — a data-fetching tool with no cross-model review process. |
| Runtime assurance | 1 | Automatic fallback mechanisms when a preferred data provider fails; no claim-verification or citation-audit layer. |
| Cross-platform portability | 2 | MCP server usable from Claude Code, Codex, and any compatible MCP client, plus a self-hostable stack and a browser demo. |
Scored on 2026-10-01. See the evaluation rubric.
Tags¶
Pipeline stages: data-acquisition
Architectural features: tool-use persistent-memory
Inputs: natural-language-query
Outputs: datasets csv-json-stata-python-exports
Data sources: fred world-bank imf eurostat un-comtrade bis statistics-canada oecd exchangerate-api coingecko chinamacro
Limitations¶
- AGPL-3.0 license requires source-sharing for hosted deployments, a heavier obligation than the catalog's typical MIT/Apache entries.
- No reported evaluation of whether natural-language queries are correctly mapped to the intended indicator series across providers.
- Free demo capped at 20 queries; self-hosting requires OpenRouter API credentials and infrastructure (Redis, Node.js, Python).