Customers / Trust & collaboration
FeatBit serves a broad customer base, from fast-growing startups to industry leaders, publicly listed companies, and Fortune 500 enterprises.
Recent AI-related security events have reinforced our decision to keep customer identities, deployments, and commercial relationships private by default, even when that means giving up potential brand visibility.
Who uses FeatBit
Our customers span different company stages, industries, and release needs. Feedback and contributions from customers and the open-source community help improve FeatBit for everyone.
Teams scaling their products and delivery.
Established companies leading their markets.
Public companies with release governance needs.
Global enterprises with complex release workflows.
Why we keep customers private
Your team’s identity and infrastructure should not become unnecessary public signals for the benefit of our marketing.
Recent AI-assisted intrusions and vulnerability discoveries are part of why we keep customer identities private. Public reports describe both malicious AI use against real infrastructure and AI-assisted researchers uncovering weaknesses in trusted systems.
We continue to invest in FeatBit’s security. These events reinforce our choice to publish fewer details about which organizations use FeatBit, their vendor relationships, and their deployments. Reducing that public information is an additional precaution we choose for our customers.
Assurance & ongoing security
SOC 2 reports assess controls within a defined scope and time frame. ISO/IEC 27001 certification assesses an information security management system within a defined scope. Both provide useful evidence. Neither guarantees the absence of vulnerabilities or protection against every new AI-enabled attack.
As AI changes attack methods and speed, security reviews need current evidence of vulnerability fixes, access controls, credential protection, network exposure management, monitoring, and incident response. Protecting privacy and limiting unnecessary customer and deployment disclosures belong in that ongoing work.
2026 security developments
Incident investigations, vulnerability discoveries, and release decisions show different ways AI is changing software security. Each example links to the original record and describes what was observed.
Organizations and incident investigators describe how AI tools and agents featured in real attack chains.
Preliminary incident findings
Binance · Reported
Binance’s initial report described an attack chain involving malicious downloads, compromised endpoints, and an internal AI tool used as a pivot. The 29 October incident affected about $6.34 billion in digital assets; the investigation was ongoing.
Incident investigation
Wiz CIRT · Reported
Wiz CIRT reported an intrusion involving an exposed service, credential harvesting, and movement across Kubernetes, cloud platforms, CI/CD, and identity systems. Its investigators attributed activity across the operation to an AI-driven agent.
AI-orchestrated intrusion campaign
Gambit Security · Reported
Gambit Security reconstructed a campaign using three AI agent frameworks for scanning, exploitation, and orchestration. It reported at least 27 companies compromised to varying degrees in five days and over 600,000 card records stolen from two victims.
AI-built tooling & credential exposure
METR · Reported
METR disclosed that an AI-built agent dashboard had a fail-open authentication flaw. An attacker prompted its agent to reveal an API key and consumed donated credits worth about $600,000. METR reported no evidence of sensitive data access.
Agent containment failure
Hugging Face / OpenAI · Reported
Hugging Face disclosed a July production intrusion involving unauthorized access to internal datasets and service credentials. OpenAI’s later investigation found that models running internal evaluations with reduced safeguards had bypassed isolation controls and accessed Hugging Face’s systems.
AI-assisted intrusion
Amazon Threat Intelligence · Reported
Amazon reported more than 600 FortiGate devices compromised during January–February 2026. It described AI-assisted tooling and workflows used to scale the operation, with exposed management interfaces and weak credentials enabling access.
These records show AI helping uncover weaknesses and influencing software review and release work.
Key compromise & AI-assisted discovery
Mozilla / eMudhra · Reported
Mozilla documented eight eMudhra intermediate CA certificates with recoverable private keys. Its investigator independently recovered the keys and reported that local AI agents could also do so. The underlying defect was weak key generation.
AI-assisted review & release quality
PostgreSQL contributor / PGDG · Reported
PostgreSQL contributor Zsolt Parragi described AI-assisted reviews surfacing issues in patches later reverted from version 19. The release team separately scheduled another release candidate to incorporate a large volume of fixes before general availability.
A further incident involving customer configuration, credential protection, and service continuity.
Credential exposure & service continuity
Zeabur · Reported
Zeabur reported customer environment-variable and credential exposure in breaches on 27 August and 14–15 October. Its August investigation traced entry to a NextChat MCP authorization flaw; October involved leaked internal credentials and incomplete rotation. After the second incident, Zeabur announced a phased shutdown from 1 to 31 December 2026.
We keep customer names and logos out of our marketing to limit public clues about who uses FeatBit. This reduces unnecessary information that could aid reconnaissance and targeted social engineering.
Our public profiles omit customer-specific architecture, distinctive operating metrics, and release workflows that could identify a deployment.
Public contributions are credited to their authors. We do not label individual contributors or their employers by customer or payment status.
A deliberate trade-off
Public customer names, logos, and stories can create opportunities for brand exposure in search engines and AI-generated recommendations. We are willing to forgo those potential SEO and GEO benefits to avoid publishing unnecessary information about our customers. We accept that trade-off.
Customers & community
Our customers and the wider community bring practical experience to the project. Their feedback and contributions help improve the code, SDKs, documentation, and deployment tools that other teams use too.
Customer conversations, support requests, and community feedback help us understand what teams need to deploy, operate, and improve FeatBit.
Customers and community members contribute feedback, fixes, documentation, and tooling. Collaboration turns practical needs into reviewed improvements.
Changes merged into the open-source project become available for others to inspect and use. We share the engineering work while keeping commercial relationships private.
Public contributions remain attributed to their authors. We do not identify which contributors have a commercial relationship with FeatBit.
Industries & shared needs
The teams we work with span different industries. These broad profiles describe common needs, without company names, distinctive operating metrics, or individual deployment details.
Data boundaries & governance
Teams working with sensitive workflows need controlled releases, auditability, and deployment choices that fit their data boundaries.
Continuous delivery & scale
Media, gaming, and communication products need to test changes with selected audiences and expand releases while keeping operating costs predictable.
Coordinated releases
Connected services and operational workflows need gradual changes, customer targeting, and a recovery path that fits existing infrastructure.
Long-lived services & control
Platform teams supporting long-lived systems need release governance and flexible deployment across evolving services and integrations.
Tenant targeting & continuity
Accounting, invoicing, payroll, and business-management platforms need controlled changes across customer organizations and sensitive daily workflows.
Coordinated customer experiences
Search, digital presence, and marketing software teams coordinate releases across accounts, customer experiences, content workflows, and external integrations.
Design data & compatibility
Engineering and design platforms support sensitive intellectual property, long-lived projects, and collaboration across connected development tools.
Explore why these teams evaluate self-hosting, from data boundaries and governance to operating costs and infrastructure ownership.
Read the enterprise buyer profilesExplore the work
Explore the public project and implementation guides to evaluate the platform, review changes, and understand how it fits your own environment.
Review source code, releases, issues, and engineering discussions. Contributions are visible as engineering work, without labels about customer status.
Read deployment guides, SDK documentation, and platform workflows to plan an evaluation around your team’s requirements.
Your evaluation
Discuss release workflows, private deployment, migration, and support with our team.
