Screenshots 1
Starwell
Summary
An AI agent that quotes a GDP figure without a citation is a liability waiting to get flagged — and every general-purpose LLM does exactly that. Starwell exists to sit between the model and the answer, pulling verified numbers from official statistical agencies so the figure your agent returns has a paper trail.
Starwell indexes datasets from 29 official sources — World Bank, Eurostat, Statistics Canada, FRED, ILO, and others — and exposes them through a single REST API and an MCP server. Every response ships with the source URL, the retrieval timestamp, the connector version, the license, and a citation block pointing back to the original release. Cross-country comparisons work in a single API call: one question, four economies, each value attributed to its own agency's methodology. The ceiling appears when your project needs sources that are not yet indexed — the vendor states counts are growing weekly, but 29 sources leaves real gaps for regional or non-English-speaking markets. Teams that hit those gaps fall back to scraping or manual data entry, which defeats the provenance guarantee entirely.
Bottom line: Starwell is the right call when your agent needs to quote inflation, GDP, or labor statistics from official sources without fabricating — it breaks down the moment your use case depends on a country or dataset not yet in the catalog.
Pricing Plans
Usage-Based- Free Tier
- 500 credits/month, 10,000 data calls/day, 6 analyses per 10 minutes, 60 saved chats, 40 analyses/day
Free
500 credits/month, 10,000 data calls/day, 6 analyses per 10 minutes, 60 saved chats
- All sources
- Cross-source up to 6 series
- Cited answers
- REST and MCP access
Starter
2,000 credits/month, 50,000 data calls/day, 12 analyses per 10 minutes, 200 saved chats
- Everything in Free
- Higher limits
Pro
10,000 credits/month, 250,000 data calls/day, 30 analyses per 10 minutes, 500 saved chats
- Everything in Starter
- Higher limits
Scale
50,000 credits/month, 1,000,000 data calls/day, 60 analyses per 10 minutes, 1,500 saved chats
- Everything in Pro
- Highest limits
View full pricing on starwell.dev →
Pricing may have changed since last verified. Check the official site for current plans.
Community Performance Report Card
No community ratings yet. Be the first to rate this tool!
Pros
Sign in to edit- Every observation includes a provenance block — source URL, retrieval timestamp, connector version, and a full citation — so your application can display or log exactly where a number came from, which means audit trails that would otherwise require manual sourcing are automatic.
- A single API call joins data across multiple official sources in one response, so building a cross-country comparison that would normally require separate agency API integrations, format normalization, and merge logic is reduced to one endpoint call.
- Provider-agnostic MCP server support means the data layer plugs into agent frameworks that speak the Model Context Protocol, so you are not rewriting integration code every time the agent runtime changes.
- Continuous verification status is attached to each series — the FRED example shows `lastVerifiedAt` timestamped to the minute — so stale or broken upstream feeds surface in the response rather than silently passing bad data downstream.
- A free tier with no card required allows developers to test real queries against the live catalog before committing budget, which means the integration risk is validated on actual data, not mocked responses.
Cons
Sign in to edit- The catalog covers 29 official sources as of the vendor's stated count; any dataset outside that list — sub-national statistics, newer emerging-market agencies, sector-specific regulators — returns nothing, and the team is back to manual sourcing with no provenance guarantee, at which point the main value proposition evaporates.
- There is no self-hosted deployment option documented anywhere on the vendor page, which means all queries and data requests route through Starwell's infrastructure; teams operating under data residency requirements or strict network egress policies cannot use this tool without an exception or an architecture workaround.
- Advanced features beyond free-tier query limits are gated behind paid access, so teams that need high-volume batch pulls or extended historical depth will hit rate or volume walls during load testing and must upgrade before production deployment.
- Teams that need proprietary, licensed, or embargoed data — IMF restricted datasets, Bloomberg terminal feeds, paid Eurostat extracts — will find the catalog does not cover those, and will need a separate data pipeline regardless of how well Starwell handles the public-domain sources; at that point, some teams consolidate onto a single data vendor that covers both, abandoning Starwell to avoid maintaining two parallel integrations.
About
- Platforms
- Web, REST API, MCP server
- API Available
- Yes
- Self-Hosted
- No
- Last Updated
- 2026-09-10T16:23:42.301Z
Best For
Who it's for
- AI agents needing non-hallucinated statistics
- Developers building data-backed applications
- Analysts requiring sourced and verified figures
- Research workflows needing official data provenance
What it does well
- Querying and aligning economic indicators across countries
- Fetching cited observations from official statistical agencies
- Generating verified reports for AI agents or analysis
- Monitoring data series with provenance tracking
Integrations
Add notes, reviews, and benchmarks so the next visitor gets a clearer picture.
Compare Starwell
Spotted incorrect or missing data? Join our community of contributors.
Sign Up to ContributeRecommended skills for this tool
Auto-curated by the AIDiveForge recommendation matrix. These skills are predicted to enhance this tool based on category, capability, and domain signals.
-
Meeting Summary Template transform 32%
Turn a raw transcript into a decision-focused recap: outcomes, owners, deadlines, open threads.
Why: category partial · caps 0/0 · domain ops
-
Standup Note Synthesizer transform 32%
Merge individual standup bullets from multiple people into a single team digest with blockers surfaced to the top.
Why: category partial · caps 0/0 · domain ops
-
Runbook Skeleton post 32%
Produce a first-draft runbook from a postmortem — detection, diagnosis, mitigation, rollback — so the next incident has a template to follow.
Why: category partial · caps 0/0 · domain ops
-
OKR Draft Critiquer post 32%
Score draft OKRs against SMART criteria and the outcome-not-output rule, with suggested rewrites for each failing key result.
Why: category partial · caps 0/0 · domain ops
Frequently Asked Questions
- Is Starwell free?
- Starwell has a permanent free tier alongside paid upgrades. You can keep using a baseline version indefinitely without paying.
- Is Starwell open source?
- No — Starwell is a closed-source tool. Source code is not publicly available.
- Does Starwell have an API?
- Yes. Starwell exposes a developer API. See the official documentation at https://starwell.dev for details.
- What platforms does Starwell support?
- Starwell is available on: Web, REST API, MCP server.
Curated lists that include this category
AI agents quoting GDP figures without citations create audit risks
Starwell sits between the model and the answer by pulling verified numbers from official statistical agencies. It indexes datasets from 29 sources including World Bank, Eurostat, Statistics Canada, FRED, and ILO, then serves them through one REST API and an MCP server. Every response includes the source URL, retrieval timestamp, connector version, license, and a citation block. Cross-country comparisons run in a single call, with each value tied to its agency’s methodology. The vendor states the catalog count is growing.
Key limits
The catalog stops at the 29 indexed sources. Any dataset outside that list returns nothing. No self-hosted option exists, so all queries route through Starwell infrastructure.
Pricing and access
Usage-based model with a free tier of 500 credits per month, 10,000 data calls per day, 6 analyses per 10 minutes, 60 saved chats, and 40 analyses per day. Platforms include web, REST API, and MCP server. Integrations cover REST API, MCP server, and OpenAPI spec.
Who it is for / who should skip it
Best for AI agents needing non-hallucinated statistics, developers building data-backed applications, analysts requiring sourced figures, and research workflows that need official data provenance. Skip it if projects require sources beyond the current 29 or demand self-hosted deployment under data-residency rules.