OSSInsight
OSSInsight is a free, open-source analytics platform from PingCAP that tracks over 10 billion public GitHub events in real time, letting developers, analysts and researchers rank, compare and explore repositories, contributors and whole technology ecosystems without an account.
What is OSSInsight?
OSSInsight is a free analytics platform, built and operated by PingCAP, that turns the public activity of GitHub into readable measurement. It ingests more than 10 billion GitHub events — stars, forks, commits, pushes, issues, pull requests and comments — drawing historical depth from GH Archive and live activity from GitHub's REST events API, then stores everything in TiDB, the distributed SQL database PingCAP also sells commercially. The site is, quite openly, both a public utility and a showcase for that database.
Five capabilities structure the product. The Data Explorer accepts a question written in ordinary English, generates the matching SQL through Chat2Query, runs it against the full event dataset and returns a visualisation. Repository analytics cover stars, forks, issues, commits, pull requests, contributors, languages and lines changed, with historical trends, response-time percentiles for issues and pull requests, commit-time heatmaps, and the geographic and company distribution of stargazers and contributors. Developer analytics read a single account's contribution history, working cadence and collaboration patterns. Organisation analytics aggregate that view across an entire GitHub org. Finally, any two repositories can be compared side by side on identical metrics.
Around these sits a library of more than 102 curated collections — technology domains such as databases, web frameworks, game engines, DevOps tooling, observability and Web3 — each ranked by stars, pull requests, issues or contributors, with monthly and historical trends reaching back to 2011. Anyone can add a collection by submitting a pull request containing a small configuration file.
Since 2026 the platform has leaned deliberately towards the AI ecosystem, with dedicated collections for AI agent frameworks, LLM tools, MCP servers, coding agents, RAG frameworks, inference engines and LLM developer tooling. It is also built to be read by machines: it publishes llms.txt and llms-full.txt, carries schema.org structured data on every page, and exposes both a documented REST API and an MCP endpoint that AI agents can call without any credential.
Everything is free, requires no account and imposes no paywall, and the source code is published under the Apache-2.0 licence.
What it does
- Analyse any public GitHub repository: stars, forks, commits, issues, pull requests and contributors
- Compare two repositories side by side on identical metrics
- Ask a question in plain English and get generated SQL plus a chart in return
- Browse more than 102 curated collections ranked by stars, pull requests, issues or contributors
- Track what is gaining momentum right now, filtered by programming language and time window
- Profile an individual developer or an entire GitHub organisation
- Query the same data programmatically through a public REST API or an MCP endpoint
When to use OSSInsight / When not to
A quick filter to help you decide if OSSInsight is the right fit.
When to use OSSInsight
- Developers and architects comparing two open-source projects before committing to a technical choice
- Data analysts and data journalists who need sourceable figures on open-source activity
- Teams tracking the AI ecosystem, from agent frameworks and MCP servers to coding assistants
- Maintainers watching the health of their own community: contributor growth, issue and pull request response times
- Builders of AI agents that need a free, unauthenticated REST or MCP endpoint for GitHub data
When not to use OSSInsight
- Anyone who needs analytics on private repositories, since only public GitHub data is covered
- Teams working on GitLab, Bitbucket or self-hosted forges, which are outside the scope entirely
- Users expecting figures that match GitHub's own counters exactly, as the two diverge by design
- Organisations that require contractual guarantees: there are no terms of service, no SLA and no DPA for this tool
- People looking for saved workspaces, alerts or team sharing, none of which exist without accounts
How to use OSSInsight
A typical end-to-end flow, from setup to results.
- Open the site directly: there is no sign-up, no login and no payment step of any kind
- Type a repository name, an organisation or a developer handle into the search box on the home page
- Read the generated dashboard: stars, forks, commits, issues, pull requests, contributors and their trends over time
- Add a comparison by clicking VS or appending the second repository as a URL parameter
- Open the Data Explorer and ask a question in plain English to get generated SQL and a chart
- Browse the collections section, pick a technology domain, then sort by the metric and time window you care about
- Check the trending page and filter it by programming language or by period, from 24 hours to three months
- For programmatic access, read the API documentation and issue a plain GET request without any key
- For AI agents, call the MCP endpoint with an action parameter such as collections, ranking, trending or compare
- To contribute a new collection, submit a pull request with a configuration file to the public repository
Pros & Cons
Pros
- Completely free with no account, no card and no usage limit on the web interface
- Open source under the Apache-2.0 licence, so the methodology can be audited rather than trusted
- Unusual data depth: more than 10 billion events with history reaching back to 2011
- Public API requires no authentication and states its rate limits openly, at 600 requests per hour per IP
- A dedicated MCP endpoint makes the dataset directly usable by AI agents, which remains rare
- Natural-language querying removes the need to write SQL against a very large dataset
- Methodological honesty: the FAQ explains the five reasons its figures diverge from GitHub's own
Cons
- No terms of service and no privacy policy of its own; the only legal page belongs to the publisher and describes a different product
- The API is labelled beta, carries no SLA and its rate limits are shared rather than configurable
- Coverage stops at public GitHub: no GitLab, no Bitbucket, no private repositories
- Published figures contradict each other across pages, from 10 billion events on the home page to 5+ billion in the documentation
- Without accounts there are no saved queries, no alerts and no team workspaces
- The Data Explorer page renders through JavaScript, so it is invisible to simple crawlers and text-only clients
- English-only interface with no localisation announced, and no contact page beyond an email address
Pricing & Plans
OSSInsight is provided entirely free of charge. There is no paid plan, no paid tier and no pricing page: the /pricing path returns a 404 response. The web interface, the curated collections, the Data Explorer and the public REST API are all accessible without an account, without a payment method and without any advertised price point, and the site's structured data explicitly declares an offer price of zero. The source code is additionally released under the Apache-2.0 licence, so the software itself carries no licence fee.
Data, GDPR & hosting
A consolidated view of how OSSInsight handles your data.
GDPR overview
OSSInsight itself makes no GDPR statement; the applicable text is PingCAP's privacy policy, last updated 4 February 2026. That policy is substantive: it grants European, UK and Swiss users rights of access, rectification, erasure, objection, portability and withdrawal of consent, relies on European Commission standard contractual clauses for transfers, and claims certification under the EU-U.S., UK Extension and Swiss-U.S. Data Privacy Frameworks, with TRUSTe as dispute resolution body. An Article 27 representative is formally designated, Bird & Bird GDPR Representative Services SRL in Brussels, alongside a separate UK representative. PingCAP undertakes to acknowledge requests within 15 business days and resolve them within 45. Two caveats matter: the policy never uses the words "GDPR compliant", and it covers the publisher's sites generally rather than this tool specifically.
Who owns the data?
There is no account on OSSInsight and no user content to own: the platform analyses public GitHub event data collected from GH Archive and GitHub's REST events API, so the underlying material belongs to GitHub and to the projects that generated it. The tool publishes no privacy policy of its own. The only legal document it links to is PingCAP's site-wide privacy policy, which declares a scope covering "our websites (our Sites)" and PingCAP's products, and which names PingCAP (US), Inc. as controller. That document's substantive clauses describe TiDB and TiDB Cloud rather than OSSInsight, so ownership questions specific to this tool are simply not addressed anywhere.
Reuse rights
Nothing on the site grants or restricts an explicit reuse right, because OSSInsight publishes no terms of service at all: the /terms, /legal and /privacy paths all return 404. What can be established is factual rather than contractual. The analysed material is public GitHub event data, and the platform's own source code is released under the Apache-2.0 licence, which permits reuse, modification and redistribution of that code. The public REST API is offered without authentication under published rate limits of 600 requests per hour per IP address and 1,000 requests per minute globally, which is a technical permission to consume the data rather than a licence covering it. Anyone republishing figures commercially should seek written confirmation, since no document defines the terms.
Data retention & training
Hosting summary
No hosting information is published for OSSInsight specifically. The applicable statement comes from PingCAP's privacy policy, which says that the publisher's website servers are located in the United States, and that its affiliates, partners, third parties and service providers operate in the United States, the European Economic Area, Canada, China and Australia, meaning personal information may be processed in any of those countries. Transfers out of the EEA, the United Kingdom and Switzerland rely on European Commission standard contractual clauses, and PingCAP claims certification under the EU-U.S., UK Extension and Swiss-U.S. Data Privacy Frameworks. One technical caveat: the domain resolves to an anycast address on Amazon's network, which identifies the nearest content delivery node rather than the location of any server, and should not be read as a hosting jurisdiction.
Things to keep in mind
Risks and trade-offs to weigh before adopting OSSInsight.
- No terms of service, no privacy policy and no contact page exist for this tool; the only legal link points to the publisher's own policy, whose body describes TiDB and TiDB Cloud rather than OSSInsight
- The data processing agreement offered by the publisher is titled for TiDB Cloud Services and does not cover this platform, and the only sub-processor named applies to another product
- The API is beta and unauthenticated: building a production dependency on it means accepting that it can change or throttle without notice
- Metrics deliberately diverge from GitHub's own counters, so quoting them without the caveat can mislead an audience
- Numbers published on the site contradict one another between pages, which invites accidental misquotation
- Ranking developers and organisations by public activity can be misread as a measure of individual worth; contribution counts reward visible work and ignore the rest
- Star counts and trend curves are easy to mistake for quality signals, and choosing a dependency on popularity alone is a well-known way to accumulate technical debt
Setup & Integrations
Technical difficulty
Effectively zero for normal use. The platform runs in a browser with no installation, no registration and no configuration; typing a repository name into the search box is the entire setup. Programmatic use is barely harder, since the public API answers plain GET requests with no key, no token and no client library required, and the MCP endpoint works the same way with a single action parameter. Self-hosting is theoretically possible because the code is open source under Apache-2.0, but the site documents no such path, so that route would demand real engineering effort.
Deployment
Integrations
Behind OSSInsight
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement OSSInsight.
Frequently asked questions
What exactly is OSSInsight?
Where does the data come from?
Is it really free?
Why do the numbers differ from what GitHub shows?
How fresh is the data?
Is there an API, and does it need a key?
What are the API rate limits?
Can I analyse my own profile or add a collection?
Should you pick OSSInsight?
OSSInsight occupies a genuinely useful position: it measures the open-source world instead of describing it, and it does so without asking for an account, a card or an email address. The combination of scale, more than 10 billion GitHub events reaching back to 2011, and openness, with the code published under Apache-2.0 and the methodology's weaknesses documented in the FAQ, makes it credible in a way that most analytics dashboards are not. For anyone choosing between two libraries, tracking the AI agent ecosystem, or needing sourceable figures for an article or a board slide, it answers quickly and for free.
The reservations are institutional rather than technical. This tool publishes no terms of service and no privacy policy of its own; the only legal page it links to belongs to its publisher and, once opened, turns out to describe a different product line. The data processing agreement available from that publisher is explicitly written for TiDB Cloud. The public API is labelled beta, has no authentication and therefore no service commitment, and its rate limits are shared across all callers. Figures also contradict each other between pages, which is a reminder that this is a free community utility rather than a contractual service.
Read in that light, the right posture is clear. Use it freely for exploration, comparison, research and reporting, where its depth is hard to match at any price. Do not place it in a production chain that needs guarantees, and do not assume that the publisher's legal documents cover your use of it. As a lens on what is actually happening in open source, measured rather than claimed, it remains one of the more valuable free tools available.
- Choosing a selection results in a full page refresh.
- Opens in a new window.