Documentation versions and changes

2026.11.02.1 · November 2, 2026

Opening BARC no longer asks you to sign in again. If you signed in to Briard with your phone code, that sign-in is enough to open BARC, however long ago it was. You still need a phone code when you sign in to Briard, and BARC still checks your workspace role and the BARC agreement. Signing out of Briard still ends your BARC session.

The troubleshooting guide's "Authentication is too old" entry is now "Phone-code check needed": it appears only when a Briard sign-in did not include a phone code.

2026.11.01.1 · November 1, 2026

The customer documentation and the public site now use the product's current display name: Briard Alert & Response Console (BARC) at first mention on a page and BARC after. This replaces the earlier name Briard Investigations (formerly KAIDAN) in active guide text, navigation, the documentation home page and the public site. The product's own screens use the same name, and quoted screen labels match them, for example BARC alerts, Open BARC alerts, Set up BARC, Accept and enable alerts and response and Alerts and response access enabled.

Existing /kaidan addresses and section links keep working. The add-on agreement keeps its original name, Briard Investigations (formerly KAIDAN), and its terms, version and digest are unchanged; the guide now says it is the same product. The October 12 video demo and its transcript are unchanged and are labeled "(formerly KAIDAN)" where they use the earlier name. Earlier release notes keep the names used at the time.

This is a naming update only. It does not change features, access, permissions, retention or source coverage. The Briard baseline in this edition's manifest is the source revision these labels were checked against; the KAIDAN baseline is unchanged.

2026.10.28.1 · October 28, 2026

The private-evidence guide now follows one incident-case journey: ask for a closer look, review the exact access, choose originals to keep, and open evidence after its storage is confirmed. Briard preserves the selected case and its Private evidence section through the existing sign-in and phone-code checks.

KAIDAN prepares exact-item requests from collection receipts and automatically transfers matching protected originals after independent approvals. Users choose the purpose and holding/retention periods; the product handles file fingerprints and approved storage settings. Failed transfers, permitted retries, role limits, legal holds and source-reported cleanup remain explicit. Normal monitoring stays metadata only.

These instructions describe the reviewed source revisions in this edition's manifest. Local reference and automated browser checks are engineering evidence; they do not establish customer-provider operation or measured human comprehension. Deployment and live verification are recorded separately. The video walkthrough illustrates an earlier access flow and is not evidence that this new journey ran.

2026.10.24.2 · October 24, 2026

The hosted documentation now includes a build identity file to help support confirm which source revision is published. Customer instructions and their reviewed product baselines remain those of edition 2026.10.24.1. A build identity does not verify a customer's setup or understanding.

2026.10.24.1 · October 24, 2026

Clarified that typing a governance answer changes a local draft and Save answer needs a confirmed result. Added the expected-device import's empty reduced templates and copyable IT request, including the rule to ask for an unknown source date instead of guessing.

Updated KAIDAN source guidance for saving one known answer, keeping unanswered sources unknown, and requesting help from an approved case administrator. Saved requests, owner responses, case labels and exact case mapping remain separate from installation, permission, approval and verified event capture. Secrets stay with the approved administrator setup.

These instructions describe the reviewed source revisions in this edition's manifest. Deployment, provider availability and human comprehension require their own evidence. Existing fleet, source, collection and video-demo limitations remain in place.

2026.10.12.4 · October 12, 2026

Expanded the Briard video demo (12 minutes, 27 seconds) with practical supporting material for AI inventory answers and a shared-workspace chapter. A fictional purpose brief is available to inspect; guidance also explains configuration records, information rules, human-review procedures, completed tests and dated decisions. File scan results do not verify the accuracy of an answer or approve a tool.

Added workspace team instructions and ordinary administrator controls to change a joined member's role or remove access. Owners and administrators can invite teammates into the same workspace. Owner access and your own membership stay protected; current membership and role govern access after changes.

The recording's invitation delivery, account joining, private upload and scan results are simulated with fictional examples. The 11 unchanged Briard chapters preserve their existing narration and footage. KAIDAN's video remains unchanged from edition 2026.10.12.1. Captions, chapter times and the exact Briard transcript match this correction.

The five customer FAQs and their homepage, documentation and KAIDAN links remain unchanged from edition 2026.10.12.3. This edition combines that guidance with the supporting-material and shared-workspace correction.

2026.10.12.3 · October 12, 2026

Added FAQs covering the data Briard receives, data protection, the information used for results, separately approved KAIDAN investigation access and starting with a small AI inventory. The public Briard homepage has the same five questions, with a link to the detailed answers. The docs navigation, documentation map and KAIDAN overview link to the shared guide.

The answers distinguish deliberately uploaded evidence-file bytes and scanning from metadata-only discovery and normal KAIDAN monitoring. A confirmed alert does not grant additional access: deeper investigation still requires an eligible case and source, bounded policy and scope, and two distinct eligible reviewers other than the requester. Approval and actual collection remain separate.

This is a guidance and static-copy update. It does not expand ingestion, permissions, retention, provider coverage or investigation capabilities. The published Briard and KAIDAN demo media, captions and transcripts remain unchanged from edition 2026.10.12.1.

2026.10.12.1 · October 12, 2026

Refreshed the Briard video demo (10 minutes, 25 seconds) and KAIDAN video demo (12 minutes, 51 seconds) for the current platform. Chapter controls, posters, English captions, and exact narration transcripts are updated together.

Briard includes updated narration for Find AI through approved work sources; Know where incident review begins.

KAIDAN includes updated narration for Check whether a tool is sending evidence; Follow the evidence path and save the proof; Understand a larger incident through its evidence path.

The demonstrations use fictional records and retain their stated simulation, approval, collection and provider boundaries. Approved connections, permissions and workspace availability still determine which sources and fleet features you can use. The recordings do not establish real provider operation or customer acceptance. The earlier September media remains available as historical material.

2026.10.11.5 · October 11, 2026

Added the signed RC9 pilot and fresh macOS credential-provisioning instructions for an approved root system job, followed by installation with the exact protected Keychain selection. RC7 and RC8 artifacts remain unchanged. Publisher trust still requires independent administrator confirmation; the package does not approve its own policy or enroll a computer.

Isolated macOS 15 ARM64 checks passed same-boot service lifecycle operations and, separately, root provisioning plus actual locked-Keychain readback by the non-root service identity. Reboot, sustained operation, native Intel Mac compatibility, central-management rollout and full fleet acceptance remain open. These limited results do not authorize broad deployment. KAIDAN's reviewed source is unchanged.

2026.10.11.4 · October 11, 2026

Clarified that Device and model inventory follows the Workplace selected above in Shadow AI Discovery. Change workplaces using that shared selector; an unavailable selection does not silently switch inventory to another workplace. Fleet availability remains subject to the approved workspace rollout, permissions and source prerequisites.

KAIDAN's reviewed source and the signed RC8 package are unchanged. macOS/Jamf service readiness remains unverified; separate diagnostic source work is not included in RC8 and does not establish native or managed-rollout acceptance.

2026.10.11.3 · October 11, 2026

Added the signed RC8 pilot and instructions for optional hardware matching, separate device-match approval and choosing an original device record before private-name lookup. Hardware checks remain off unless explicitly approved; unavailable or conflicting identity stays visible as a coverage gap. RC7 and its policy remain available unchanged, but RC7 cannot collect the optional hardware check.

KAIDAN's reviewed source is unchanged. Native service, independent provider correlation and full fleet acceptance remain separate from source implementation and package signing. Existing feature, permission and deployment prerequisites still apply.

2026.10.11.2 · October 11, 2026

Added the public device collector deployment and maintenance guide, including independent publisher approval, protected per-device credentials, Intune/Linux installation recipes, update, rollback, revocation and removal. The fleet capability matrix now distinguishes provider licensing and visibility from Windows, Linux, macOS, model-file, WSL/container and browser limits.

macOS/Jamf rollout remains blocked pending native service verification. Isolated Keychain readback passed with the corrected setup, but service readiness was not confirmed. These instructions do not establish managed rollout, boot or fleet acceptance. KAIDAN source behavior and the signed RC7 artifacts are unchanged.

2026.10.11.1 · October 11, 2026

Added the signed collector RC7 pilot package and public publisher-policy instructions to the fleet guide. Administrators must independently confirm the publisher fingerprint and provision trust through their management channel before installation. The release includes fixed, non-sensitive installer failure stages for support. Signing does not establish fleet acceptance or enable collection.

KAIDAN main remains unchanged at the reviewed baseline. Provider licensing, operating-system acceptance, approved enrollment and customer rollout remain separate requirements.

2026.10.09.2 · October 9, 2026

Added instructions for choosing an existing expected-device group and assigning supported central network evidence to workplaces. The fleet guide explains missing group members, changed-register review, exact provider-device matching, effective dates and unresolved gaps. These remain conditional fleet features; selecting a group does not change collection access, and assigning network contacts does not prove device coverage or AI use.

KAIDAN source behavior is unchanged. This edition does not establish new provider, native fleet or hosted customer acceptance.

2026.10.09.1 · October 9, 2026

This edition adds administrator guidance for the expanded Briard fleet workflow and clarifies its relationship to existing discovery, governed records and KAIDAN. Expanded inventory, lookup, provider-control and Jira features are separately enabled and off by default. Provider permissions, licensing, approved distribution and deployment validation still determine availability. Full real-provider, hosted customer-workflow and long-running fleet acceptance are not established by this documentation edition.

Briard-AI

• Added Fleet inventory and coverage for separate sites, including medical organizations and restricted devices.
• Explained authoritative expected-device registers, automatic management-roster refresh, unknown coverage percentages, identity conflicts, source age and capability-specific gaps.
• Added guided local runtime, model-folder, software and nested-environment setup; central deployment prerequisites; and the requirement for approved signed machine distribution.
• Distinguished automatic current discovery records from reviewed governance records, installation/model availability from actual use, and browser/network/cloud signals from device coverage.
• Documented separately granted, purpose-bound device lookup with fresh MFA, expiring mappings and audit.
• Added scoped policy review, task assignment, explicitly approved Jira notifications and delivery, independent control-effect review, and pending KAIDAN handoff drafts.
• Updated Browser Check guidance for conditional store availability, administrator restrictions, user consent, offline pending notices and workspace boundaries.

KAIDAN

Clarified how Briard inventory coverage and pending handoff drafts relate to KAIDAN source proof. This edition does not announce a new KAIDAN source capability or convert a discovery record, provider receipt or ticket into an investigation conclusion. Dated recordings, evidence, agreements and commercial boundaries remain unchanged.

2026.10.08.1 · October 8, 2026

This edition covers the Briard-AI changes merged through October 8 and the KAIDAN changes merged through October 7. The documentation version identifies this reviewed set of guides; individual source availability still depends on your workspace, permissions and deployment.

Briard-AI

• Updated navigation for Home, My AI tools, What needs me, More and the learning pages.
• Added the Shadow AI Discovery guide, covering workplace selection, guided connections, administrator handoff, approved-account reuse, first-check evidence, human confirmation and coverage limits.
• Documented Google Cloud billing/API prerequisites, restricted provider identities, Salesforce installation before approval and source repair/removal.
• Added the Home AI incident alerts summary and the distinction between a failed summary, no open alerts and unmonitored tools.
• Clarified the boundary between agent governance records and separately enabled Guardian runtime controls.
• Added policy approval/publication, reading-campaign status, tool reviews, agent permissions and the incident-to-policy follow-up workflow to the user guide.

KAIDAN

• Updated guidance for the refreshed console and its preserved investigation, evidence and access workflows.
• Separated the available hosted-console path from an approved customer-controlled deployment.
• Added Check a tool link, source health, event proof, missing-data repair and gap handling.
• Clarified current priority explanations, historical assessment versions and a fictional first exercise covering all four priorities.
• Distinguished browser/app discovery from proved action capture and KAIDAN intake.

Dated material

The September 14 videos, captions and transcripts remain the record of that demonstration. Their pages now point to the current written instructions. Dated trust and recovery evidence retains its original date and scope; this documentation update does not create new provider receipts, certifications, human approvals or customer acceptance.

How version history works

The footer identifies the current documentation edition. Each edition is kept in Git with its change notes, exact Briard-AI and KAIDAN source revisions, and an immutable docs/vYYYY.MM.DD.N tag created after publication is verified. Product changes that affect customer steps require a documentation review in the same release process. Corrections receive a new edition; prior editions are not overwritten.

The normal guide URLs always serve the current edition. Earlier editions are retained as tagged source snapshots rather than separate browsing sites. Contact support@briard-ai.app with the documentation version and product page if you need an earlier guide.

Documentation was already tracked in Git before edition 2026.10.08.1. That edition is the first explicit, jointly reviewed documentation baseline; it does not invent version numbers for earlier product releases.