Feedly
Feedly is an AI intelligence platform that continuously scans 10,000+ sources to surface cyber threats, market signals and news, then turns them into cited answers, dashboards, reports and STIX feeds for security, research and strategy teams.
What is Feedly?
Feedly is an AI intelligence platform run by feedly, Inc. from California. One engine feeds three products: Threat Intelligence for cyber teams, Market Intelligence for strategy and innovation teams, and the News Reader, the long-running reading app the company is best known for. The homepage sums the promise up as the fastest way to track emerging threats with AI, positioning Feedly as the platform CTI teams use to track threats, contextualise what matters and deliver tailored intel in minutes.
The machinery behind it rests on three pillars. Intel Agents collect continuously from more than 10,000 vetted clear web and dark web sources. Ask AI researches imminent threats and drafts deliverables. Integrations then ship the result onward. Over 1,000 AI models extract entities, identify incidents and tag TTPs, feeding a Real-Time Threat Graph that links more than 10 million articles to 979 threat actors, 300,000+ CVEs, 12,000+ malware families, 800 TTPs and 681 million IoCs. Every AI answer carries deep citations back to the original article, a design choice Feedly summarises as keeping the analyst in control.
Market Intelligence reuses the same foundations for competitive work: AI Feeds, an Emerging Trends Dashboard, 360-degree Insights Cards, AI Actions that export facts into spreadsheets, and Automated Newsletters. Feedly claims three to five times faster intelligence gathering there. The News Reader covers news sites, blogs, trade publications, email newsletters and Google News feeds, with Boards for notes and highlights, cross-source search, and iOS and Android apps.
A public REST API on developers.feedly.com exposes Feed, Search and STIX endpoints plus a Threat Graph MCP server for AI assistants. Feedly reports 380 CTI teams including half of the Fortune 10, and 400+ security teams on its pricing page. The domain has been captured by the Wayback Machine since 2005, which makes this an unusually old product in the AI directory. The main friction is commercial: neither enterprise product publishes a price.
What it does
- Collect intelligence continuously from over 10,000 vetted clear web and dark web sources
- Let more than 1,000 AI models extract entities, flag incidents and tag TTPs automatically
- Ask questions in plain language and get answers that cite the article they came from
- Build briefings, reports and automated newsletters shaped for each group of stakeholders
- Push enriched intel in STIX 2.1 format into an existing security stack
- Save, annotate and highlight articles on shared Boards for team research
- Extract facts from articles into tables, charts or spreadsheets with AI Actions
When to use Feedly / When not to
A quick filter to help you decide if Feedly is the right fit.
When to use Feedly
- Cyber threat intelligence teams that must watch CVEs, threat actors, TTPs and IoCs without missing the first hours of a campaign
- Security engineers who want enriched intel pushed straight into OpenCTI, Cortex XSOAR, MISP, Anomali or Microsoft Sentinel in STIX format
- Strategy, innovation and competitive intelligence analysts tracking rivals, weak signals and emerging industry trends
- Anyone whose job ends with a deliverable: briefings, newsletters and stakeholder reports built from sourced, citable material
- Individual readers and researchers who want news sites, blogs, trade publications, email newsletters and Google News gathered in one free reader
When not to use Feedly
- Buyers who need a published price before talking to sales: both Threat Intelligence tiers are quote-only, and Market Intelligence is too
- Organisations bound to keep data inside the European Union, since production hosting is exclusively in the United States
- Teams that require a contractual guarantee their content will never reach an AI model, as no training opt-out is documented anywhere
- Users who need a documented non-English interface, because no list of supported interface languages is published
- Companies unwilling to accept mandatory individual arbitration, a class action waiver and auto-renewal without pro-rated refunds
How to use Feedly
A typical end-to-end flow, from setup to results.
- Open a free account from feedly.com, no credit card required, or start a Market Intelligence trial from the request-trial page
- Follow your first sources: news sites, blogs, trade publications, email newsletters, Google News, Reddit, Mastodon, Bluesky or X
- Create AI Feeds so the models filter the flow down to the vulnerabilities, actors, companies or trends you actually track
- Read the Insights Cards on a CVE, threat actor or malware family to get the context behind an alert
- Query the corpus with Ask AI and check each answer against the citations it returns
- Save what matters to a Board, annotate it and highlight the passages your team will reuse
- Assemble a briefing with Report Builder or schedule an Automated Newsletter for your stakeholders
- For Threat Intelligence, take the day-one kickoff call with your dedicated advisor to tune the trial to your intelligence requirements
- Connect the no-code integrations to OpenCTI, Cortex XSOAR or MISP, and add Slack or Microsoft Teams for distribution
- Generate an API access token on developers.feedly.com to pull articles, searches or STIX 2.1 objects into your own tooling
Pros & Cons
Pros
- Every AI answer is traceable to its source article, which matters when the output becomes a briefing
- Unusual graph depth: 979 threat actors, 300,000+ CVEs, 12,000+ malware families, 800 TTPs and 681 million IoCs
- Twenty-six named integrations covering the main TIP and SOAR platforms, plus Slack, Teams and Zapier
- Legal documentation is remarkably complete: DPA, SCCs, UK and Swiss addenda, SLA, dated sub-processor list, detailed security practices
- Public REST API with STIX 2.1 output and an MCP server, documented and openly readable
- A dedicated threat intelligence advisor is included from the first paid tier onward
- Free trial without a credit card on Market Intelligence and a permanent free tier on the News Reader
Cons
- Neither Threat Intelligence tier shows a price: both send you to a Request pricing form, and Market Intelligence does the same
- News Reader prices are absent from any public page and only surface through the application's catalogue endpoint, varying by country
- No documented way to exclude your data from model training, while OpenAI and Anthropic are both listed as AI sub-processors
- Production hosting is United States only, with no European data residency option and no Article 27 EU representative
- The privacy policy still carries a January 2024 effective date while the sub-processor list is dated June 2026
- The postal address differs between documents: Palo Alto in the terms of use, Redwood City in the DPA and the enterprise agreement
- Mandatory individual arbitration, a class action waiver, automatic renewal and no pro-rated refund on cancellation
Pricing & Plans
A permanent free plan is available on the News Reader, and Market Intelligence can be trialled at no cost without a credit card. The cheapest paid entry point observed is the News Reader Pro plan at 8.99 EUR per month, or 83.88 EUR per year; Pro+ is offered at 105.79 EUR per year. Those amounts come from Feedly's public catalogue endpoint and are served in local currency according to the visitor's location, so a United States visitor sees a dollar price instead. Both Threat Intelligence tiers and Market Intelligence are quoted on request only.
- permanent free tier
- no credit card required
- 8.99 EUR per month
- or 83.88 EUR billed yearly
- 105.79 EUR per year
- discounted from 145.21 EUR
- 1000+ TI AI models
- CVE
- threat actor and malware Insights Cards
- 50 AI Feeds
- 5 newsletter templates
- Ask AI and Report Builder
- dedicated TI advisor
- whole team access
- everything in Standard
- plus Vulnerability
- TTP and Cyberattack Intel Agents
- 100 AI Feeds
- 10 newsletter templates
- no-code OpenCTI
- XSOAR and MISP integrations
- Slack and Microsoft Teams
- Third Party Intelligence
- Dark Web Agents
- Commercial Access
- free trial without a credit card
- pricing on request
Data, GDPR & hosting
A consolidated view of how Feedly handles your data.
GDPR overview
Implementation is documented rather than merely claimed. Feedly publishes a Data Processing Agreement incorporating the Standard Contractual Clauses with preferred elections, a UK addendum and a Swiss addendum, and defines GDPR as Regulation 2016/679 together with the UK GDPR and the Data Protection Act 2018. The Irish Data Protection Commission is named as competent supervisory authority under SCC clause 13. The privacy policy lists the eight rights available to EU users and the right to complain to a supervisory authority, with privacy@feedly.com as the contact. A dated, named sub-processor list serves as Annex III, with objection possible within ten days. Two gaps remain: no Article 27 EU representative is designated, no DPO is named, and hosting is entirely in the United States.
Who owns the data?
Under the published Data Processing Agreement the customer is the data controller and Feedly acts only as processor. On the consumer side Feedly states plainly that you should own and control all your personal information. Account holders can update, correct or delete their data at any time from the interface or by writing to privacy@feedly.com, and customers remain responsible for deleting their own content. The Information Security Practices page adds two safeguards: customer data is not used for any purpose other than the one it was collected for or the provision of the service, and Feedly staff cannot log into or view customer accounts once advanced privacy is enabled.
Reuse rights
Feedly grants a limited, non-exclusive, non-transferable, non-sublicensable and revocable licence to install and run the app on a single device, strictly for personal or internal business purposes. Reuse beyond that is expressly forbidden: subscribers may not license, sell, rent, lease, transfer, assign, reproduce, distribute, host or otherwise commercially exploit any part of the service, and the Feedly name, logos and trademarks cannot be used without permission. One clause deserves attention: any idea, suggestion or proposal sent as Feedback hands Feedly a fully paid, royalty-free, perpetual, irrevocable, worldwide, non-exclusive and fully sublicensable licence to exploit it commercially or otherwise. Content pulled through integrated third-party tools stays governed by those providers' own terms, the privacy policy citing the YouTube API as an example.
Data retention & training
Hosting summary
Production data centres are operated by Equinix and located in the United States: two separate, distant sites, one in California and one in Virginia, with a warm standby in Virginia running synchronous replication. Feedly states that these data centres comply with SOC 1 Type II, SOC 2 Type II, PCI DSS and NIST 800-53 and are ISO 27001 certified, the DPA adding FISMA to that list. Production systems and data are fully segregated from the company's internal business systems, physical and logical access is restricted through multi-factor, IP-controlled controls, traffic is encrypted over HTTPS and passwords are stored only as hashes. Declared hosting sub-processors are Alphabet and Equinix, with Cloudflare providing cloud services; DNS resolves to a Cloudflare anycast address. Almost all named sub-processors also process in the United States, the single exception found being DropContact in France. No European residency option is offered.
Things to keep in mind
Risks and trade-offs to weigh before adopting Feedly.
- Cited answers feel authoritative, and that is exactly when people stop opening the source: the citation is an invitation to verify, not a substitute for it
- Automating briefings and newsletters can quietly transfer editorial judgement to a filter, so a team stops noticing what its AI Feeds never surfaced
- Continuous threat alerting produces fatigue; after a few weeks the reflex to skim rather than read is a real operational risk
- Nothing published states whether your content trains AI models, while OpenAI and Anthropic sit on the sub-processor list under Artificial Intelligence
- Any feedback, idea or proposal you send grants Feedly a perpetual, irrevocable, worldwide and sublicensable licence to exploit it
- United States-only hosting with no Article 27 EU representative may not fit an organisation with European data residency obligations
- Subscriptions renew automatically at the then-current price, disputes go to individual arbitration, and cancellation brings no pro-rated refund
Setup & Integrations
Technical difficulty
Low to start, moderate to deploy. A free account takes minutes, needs no credit card, and following sources or creating AI Feeds requires no technical skill. Team rollout adds SAML, Google or Azure single sign-on with domain restriction. The no-code OpenCTI, XSOAR and MISP connectors on the Advanced tier avoid custom development, while API work needs a generated access token and some familiarity with REST and STIX 2.1. Threat Intelligence customers get a dedicated advisor and a day-one kickoff call, and documentation is available at docs.feedly.com and developers.feedly.com.
Deployment
Apps stores
Integrations
Behind Feedly
Social
Resources
All the official URLs gathered for verification and reference.
Frequently asked questions
Does Feedly publish its prices?
Is there a free plan or a free trial?
Is there an API?
Where is the data hosted?
Is a Data Processing Agreement available?
Is customer data used to train AI models?
What is the minimum age to use Feedly?
How long is data kept after cancellation?
Which apps and extensions exist?
Should you pick Feedly?
Feedly is one of the rare AI tools in this directory with a long history behind it, and that shows. The Threat Graph is genuinely deep, the integration list covers the platforms security teams already run, and the insistence on citing every AI answer back to its source article is the right instinct for work that ends in a briefing someone will act on. The legal shelf is equally serious: a real DPA with Standard Contractual Clauses, UK and Swiss addenda, a dated and named sub-processor list, and a security page that commits to specific numbers rather than vague assurances.
The friction is commercial and informational. Neither enterprise product shows a price, so evaluating Feedly against a competitor means booking a call first. The News Reader prices exist but hide inside the application, served in local currency, which makes them awkward to quote. More substantially, nothing on the site says whether customer content is used to train models, and that silence sits uneasily beside a sub-processor list naming both OpenAI and Anthropic under the purpose Artificial Intelligence. European buyers face a second constraint: production hosting is United States only, with no residency option and no Article 27 representative.
The fit is therefore clear. For a CTI or competitive intelligence team that wants breadth of collection, traceable AI output and a delivery path into its existing stack, Feedly is a strong candidate and the free reader makes the interface easy to sample first. For a buyer who needs published pricing, European hosting or a written commitment on model training, the answers are not on the website, and they should be asked for in writing before signing.
- Choosing a selection results in a full page refresh.
- Opens in a new window.