DFIR-IRIS vs Grail

DFIR-IRIS is an open-source incident response platform for teams conducting investigations in a shared workspace. Case records organize assets, indicators of compromise, tasks, evidence and notes with an audit trail. Alerts from SIEM, EDR, MISP, IntelOwl or REST endpoints can be collected, triaged and promoted into cases. Evidence can be checked with SHA-256, MD5 or SHA-1, while access restrictions and logs cover downloads, views and changes. War rooms provide real-time chat, threads, room-specific teams and slash commands that record decisions and case actions. Configurable Jinja templates generate PDF or DOCX executive summaries, technical reports or court-ready documents. The project lists modules including IrisVT, IrisMISP, IrisCheck and IrisWebHooks, along with a REST API and 162 hook binding points. Permissions include deny-all, read-only and full-access case controls. Docker and Docker Compose are recommended; Linux and macOS are officially supported, while Windows may need Dockerfile changes for file storage paths. Documentation identifies v3.0.0-beta.1 as the current v3 release and says beta versions are not production-ready. The project advises restricting access and not exposing IRIS to the Internet.

Grail is a cloud incident-response pipeline from Grail Forensics, Inc. Six AI agents handle intake, endpoint evidence collection, analysis, validation, and reporting. The workflow uses Velociraptor on demand, reconstructs timelines from process, network, registry, and filesystem events, and maps suspicious behavior to MITRE ATT&CK. A validation agent checks findings against raw evidence and rejects claims not anchored to hashed artifacts. Reports contain an executive summary, technical appendix, and evidence index, with claims linked to source artifacts, lines, and SHA-256 hashes and chain of custody sealed end to end. Grail lists alert intake from Sentinel, Splunk, CrowdStrike, SentinelOne, and Elastic, with outputs to Jira, ServiceNow, carrier claims APIs, and custom webhooks. Evidence is tenant-isolated, encrypted with a customer-held KMS key, purged after delivery, and not used to train models. Pricing is capped per case and available on request.