SignalForge · A Kaji Labs product
Access by request
One reporting model for every system you run.
SignalForge syncs cloud, HR, ERP, identity, AI-tool and on-prem data into one model, with dashboards, alerts and AI insights on top. 13 connectors available today, self-hosted or run by us.
Core hierarchy
Six concepts, one pipeline.
System template
A YAML blueprint for a system type: endpoints, auth, paging, field mappings and schemas. Define your own for any REST API.
Data collection
Connectors are configuration, not code.
Templates say what to collect; the platform handles auth, paging, rate limits, retries and change detection.
Delta sync
Incremental sync on updated-at fields, row versions or a custom query, with lookback windows and a scheduled full reconcile.
On-prem, outbound only
A worker agent inside your network connects outbound over WebSocket and collects LDAP and SQL data. No inbound rules.
Classified failures
A failed sync reports one of twelve categories, such as credential rejected or rate limited, with guidance instead of stack traces.
Query & reporting
A query engine, not string-built SQL.
Reports compile on the server from a validated expression tree, with field-level security applied in the compiler.
Rich aggregations
Count, sum, average, median and conditional variants, with group-by, top-N, date bucketing and relative ranges like now-7d.
Cross-system identity joins
Join records about the same person across identity sources, such as Active Directory and Entra ID, in one report.
Exports and schedules
CSV, XLSX and PDF, streamed for large results, plus cron-scheduled email delivery with per-schedule error tracking.
Compare, track & share
Numbers with their context.
See a figure against the period before, against its goal and against what happened that week, then share it without handing out an account.
Compare periods
Any report against the period before or the same period last year, from one switch in the query builder.
Goals, notes and history
A goal on a card shows how far there is to go. Notes pin events to a day. Every saved change is kept and can be put back.
Share links and wall screens
A link shows a snapshot of what its maker can see, never a live query. It expires, counts its views and can be revoked.
Automation
Findings that route themselves.
Alerts, webhooks and tickets make sure someone acts on what a report finds.
Threshold alerts
Rules on any report field, with Info, Warning and Critical severity and cooldowns that prevent alert storms.
Signed webhooks
HMAC-SHA256-signed JSON or ArcSight CEF payloads for your SIEM, with backoff retries and a per-delivery log.
Tickets and notifications
Jira or ServiceNow tickets from security findings, kept in sync, and in-app notifications when a sync completes or an alert fires.
AI (preview)
AI that explains, with guardrails.
The Forge Assistant answers questions about your data, draws the chart and proposes reports, designed on the assumption that the model is untrusted. The assistant is included from Pro; the MCP server for your own AI tools is part of Enterprise.
Bring your own model
Anthropic, any OpenAI-compatible endpoint (including local models via LM Studio, Ollama or vLLM) or Azure OpenAI.
Permission-scoped context
Context comes from running each report as the requesting user. Reports they can't read are excluded, and the summary says so.
Proposes, never saves
The assistant drafts a report or a dashboard change; a person applies it. Connected AI agents (MCP) get only the tools you allow.
Security analytics
Identity analytics across every system.
A licensable module correlates persons, accounts and entitlements across systems into one model, then analyzes it.
Risk scoring & SoD
Explainable per-principal risk scores with history, and segregation-of-duties rules with a violation workflow.
Access reviews & lifecycle
Certification campaigns with reviewers, delegation and decision trails, plus joiner-mover-leaver detection.
Compliance & posture
Controls mapped to 12 public frameworks, evaluated against synced data, with trends, evidence packages and cloud posture checks.
Attack-path graph
A graph of identities and resources: paths to Tier-0 assets, blast radius and role-mining recommendations.
Governance & operations
Enterprise controls, built in.
What a security review asks about, in the product today.
Access control to the field
Four layers: authentication, feature permissions, per-resource ACLs, and column and row security that fails closed.
SSO, MFA and SCIM
Entra ID or any SAML 2.0 or OIDC provider, optional enforced SSO, TOTP and FIDO2 MFA, and SCIM 2.0 provisioning.
Audit and operations
Tamper-evident audit logging, per-source archive and purge rules, and background jobs that run exactly once across replicas.
Signed exports (Enterprise)
Export a report or dashboard with a signature an auditor or a lawyer can check without an account: the exact file, who exported it, when, and how current the data was. Your own key or one generated for you.
See your own systems in SignalForge.
Nothing is charged until you choose a paid plan.
Want to walk through it with us first? Talk to us.