Cortex
Self-hosted homelab log intelligence over MCP, CLI, and REST with SQLite/FTS.
It collects logs and operational evidence, stores them in SQLite with FTS5 search, and exposes one shared intelligence layer through CLI, REST, MCP, and a bundled browser workspace.
Cortex began as a syslog receiver. It now covers network logs, Docker, managed files, OpenTelemetry logs, metrics, and traces, host heartbeats, fleet inventory, shell and agent activity, and Claude, Codex, Gemini CLI, and Antigravity transcripts. It correlates those sources into timelines, incidents, and an evidence-backed topology graph without making the graph a second source of truth.
At a glance
| Area | What Cortex provides |
|---|
| Ingest | UDP/TCP syslog, OTLP/HTTP logs, metrics, and traces, Docker logs and events, managed file tails, host heartbeats, AI transcripts, shell history, agent command records, and fleet inventory |
| Storage | SQLite in WAL mode, FTS5 full-text search, bounded metadata, retention, storage budgets, maintenance jobs, checkpoints, and 63 sequential schema migrations |
| Investigation | Search, filtering, context, timelines, patterns, anomaly comparison, cross-source correlation, recurring error signatures, deterministic incident bundles, and graph explanations |
| Fleet intelligence | SSH and API inventory collectors, host state, service topology, container and route relationships, redacted evidence, and rebuildable graph projections |
| AI operations | Claude, Codex, Gemini CLI, and Antigravity session indexing; skill, MCP, and hook event extraction where each provider exposes them; incident clustering; and guarded local LLM assessments |
| Interfaces | Native CLI, one action-dispatched MCP tool, authenticated REST APIs, MCP prompts and resources, an MCP Apps search widget, and a bundled investigation workspace |
| Operations | Setup and repair, diagnostics, Compose control, backup, integrity checks, WAL checkpoints, vacuum, update workflows, agents, and health endpoints |
IMPORTANT
Cortex is designed for a trusted homelab or small private fleet. It is not a clustered log warehouse, a general-purpose SIEM, or a safe place to expose unauthenticated administrative surfaces to the public internet.
Contents
Quick start
Install the CLI
The first-class guided path is:
npx skills add dinglebear-ai/cortex --skill install-cortex
Start a fresh agent task and invoke $install-cortex. It chooses server or client-only role, delegates to the canonical installer/setup engine, configures auth/network/storage, and verifies the live client connection.
The npm launcher remains the fastest manual path for local CLI and stdio MCP use:
npx -y @dinglebear/cortex --help
npx -y @dinglebear/cortex mcp
Install it permanently with:
npm install --global @dinglebear/cortex
cortex --version
The launcher requires Node.js 18 or newer. It downloads a checksum-verified native release binary for Linux x64, Linux ARM64, macOS ARM64, or Windows x64.
Build from source with the current stable Rust toolchain:
git clone https://github.com/dinglebear-ai/cortex.git
cd cortex
mise install
just build
./.cache/cargo/debug/cortex --version
Start a local server
The full daemon starts UDP and TCP syslog receivers plus the shared HTTP server. Use separate MCP and REST tokens:
mkdir -p "$HOME/.cortex/data"
export CORTEX_DB_PATH="$HOME/.cortex/data/cortex.db"
export CORTEX_TOKEN="$(openssl rand -hex 32)"
export CORTEX_API_TOKEN="$(openssl rand -hex 32)"
cortex serve mcp
Defaults:
- Syslog:
0.0.0.0:1514 over UDP and TCP
- HTTP:
127.0.0.1:3100
- MCP:
http://127.0.0.1:3100/mcp
- REST:
http://127.0.0.1:3100/api/*
- Investigation workspace:
http://127.0.0.1:3100/app
Verify it from another terminal:
curl -fsS http://127.0.0.1:3100/health
logger -n 127.0.0.1 -P 1514 --tcp "cortex quickstart from $(hostname)"
export CORTEX_API_TOKEN="the-same-api-token"
cortex tail --limit 10
For a managed local deployment, cortex setup repair creates or repairs the Cortex home, Compose assets, data paths, and missing 64-character MCP and REST tokens without replacing existing token values.
Connect an MCP client
Query-only stdio mode reads the configured local database and starts no network listeners:
{
"mcpServers": {
"cortex": {
"command": "npx",
"args": ["-y", "cortex-rmcp", "mcp"],
"env": {
"CORTEX_DB_PATH": "/absolute/path/to/cortex.db"
}
}
}
}
Streamable HTTP mode connects to the persistent daemon:
{
"mcpServers": {
"cortex": {
"url": "http://127.0.0.1:3100/mcp",
"headers": {
"Authorization": "Bearer your-cortex-token"
}
}
}
}
A useful first call is:
Then narrow the investigation with tail, errors, search, timeline, or context before using broader analysis operations.
How Cortex is built
Cortex is one Rust binary with multiple operating modes. The same application and service layer backs the CLI, REST handlers, and MCP handlers, so validation, limits, identity resolution, redaction, and business rules do not belong to one transport alone.
INGESTION
Syslog UDP/TCP OTLP signals Docker agent / pull
Managed file tails Heartbeats Claude / Codex / Gemini / Antigravity
Shell history Agent commands Fleet inventory
\ | /
\ | /
+---- bounded parsing and enrichment ----+
|
scrub, normalize, batch
|
SQLite WAL + FTS5
|
+----------------+----------------+
| |
authoritative records derived accelerators
logs, heartbeats, rollups, signatures,
inventory, sessions graph projections
| |
+----------------+----------------+
|
shared service layer
|
CLI REST MCP Web workspace
The daemon supervises its receivers and background services with cooperative cancellation. Shutdown drains HTTP requests, gives maintenance tasks 10 seconds to finish before abort-and-join, gives ingest 5 seconds to flush, and then attempts a WAL checkpoint. Already-running blocking SQLite calls cannot be cancelled by Tokio.
Background services include:
- Retention and storage-budget enforcement
- WAL and FTS maintenance
- Docker ingest supervision
- File-tail supervision
- Error-signature scanning
- Notification evaluation, dispatch, and digest scheduling
- Inventory refresh and backfill
- Graph projection refresh
- AI-session and timeline rollups
- Database optimization and maintenance jobs
Heavy analytical reads and maintenance jobs have separate concurrency controls so one expensive investigation cannot starve the ingest path.
Ingestion
All log-like sources are normalized into the same durable log model, enriched where safe, scrubbed where configured, and written through bounded batch paths.
Syslog over UDP and TCP
Cortex listens on the same configurable port for UDP and TCP syslog. It parses common RFC 3164 and RFC 5424 shapes, preserves the raw frame, records sender identity, normalizes severity and facility, and enriches known application formats.
Relevant defaults:
- Bind:
0.0.0.0:1514
- Maximum message: 8 KiB
- Maximum concurrent TCP connections: 512
- TCP idle timeout: 300 seconds
- Writer batch: 100 records or 500 ms
- Write queue capacity: 10,000 records
Syslog has no application-layer authentication. Restrict senders with network controls and CORTEX_ALLOWED_SOURCE_CIDRS when the listener is reachable beyond a trusted network.
Built-in enrichment recognizes useful signals from AdGuard, Authelia, Docker lifecycle events, fail2ban, Linux kernel and OOM events, SWAG, reverse-proxy logs, and host-local Cortex Docker agent metadata. Source gates can restrict enrichment that would otherwise trust a marker inside an unauthenticated syslog body.
OpenTelemetry ingestion
Cortex accepts protobuf OTLP/HTTP export requests on the shared HTTP listener. Logs use POST /v1/logs with a 4 MiB body limit and flow through the normal writer. Metrics use POST /v1/metrics and traces use POST /v1/traces, each with an 8 MiB body limit and dedicated storage. Metric requests admit at most 5,000 points and also enforce a bounded normalized payload budget.
Current OTLP scope is intentionally narrow:
- Logs, traces, and metrics over HTTP/protobuf are supported.
- OTLP/gRPC is not implemented.
All three OTLP endpoints authenticate with CORTEX_TOKEN — the same static MCP bearer token that guards POST /mcp, read from the managed ~/.cortex/.env on a deployed host. It is not CORTEX_API_TOKEN (REST /api/*) and not CORTEX_API_ADMIN_TOKEN. Loopback and trusted-gateway policies skip the check. An OAuth-only deployment with no static token denies OTLP outright, because machine exporters have no OAuth flow — so a non-loopback OAuth-only OTLP exposure is rejected at startup unless CORTEX_TOKEN is set.
Docker logs and events
Cortex supports two Docker collection paths:
- Host-local agent, the preferred multi-host path. The host-local cortex agent reads the local Docker socket, converts logs and lifecycle events into bounded records, and forwards them to the server without changing Docker's daemon logging driver.
- Central pull compatibility mode, an optional server-side collector for explicitly configured Docker Engine or docker-socket-proxy HTTP endpoints. It records per-container checkpoints and reconnects with bounded exponential backoff.
Central pull is disabled by default. The CORTEX_DOCKER_HOSTS shorthand expands hosts into insecure http://host:2375 endpoints and should only be used on a tightly controlled private network. A hosts file supports explicit base URLs and safer endpoint configuration.
Managed file tails
Managed file-tail sources are persisted in a registry and supervised by the daemon. Add, remove, list, and inspect sources through the CLI, REST, or the file_tails MCP admin action.
The path policy rejects unsafe targets, including paths outside configured roots, symlink escapes, non-regular files, and sensitive mounts. Container deployments expose an explicit read-only file-tail root rather than the entire host filesystem.
Host heartbeats
The host agent can post bounded JSON snapshots to POST /v1/heartbeats. Heartbeats include host state such as load, memory, disks, networking, processes, and container summaries. They power host_state, fleet_state, and correlate_state.
Heartbeat request bodies are capped at 256 KiB. Heartbeat data has short operational retention separate from the main log-retention policy.
AI transcripts
Cortex indexes local and forwarded transcript data from:
- Claude Code projects under
~/.claude/projects
- Codex sessions and worktrees under
~/.codex/sessions and ~/.codex/worktrees
- Gemini CLI chat data under
~/.gemini/tmp
- Antigravity desktop's redacted transcript projections under
~/.gemini/antigravity/brain/<session>/.system_generated/logs/transcript.jsonl
- Antigravity CLI's redacted transcript projections under
~/.gemini/antigravity-cli/brain/<session>/.system_generated/logs/transcript.jsonl
The Antigravity adapter reads only the narrow redacted JSONL projections above; it does not scan conversation databases or other brain artifacts. The scanner supports incremental checkpoints, parse-error records, bounded chunks, broad-path rejection, and safe recovery from changed files. It extracts normalized transcript rows plus dedicated skill, MCP tool-call, and hook events where those lanes are represented by the provider schema, and reports unsupported lanes honestly.
A satellite agent can send already-parsed records to POST /v1/ai-transcripts, which prevents transcript collection from depending on the database living on the same host as the AI client.
Shell and agent activity
Satellite agents can forward additional operational evidence to the shared server:
POST /v1/agent-commands for deduplicated agent command-spool recordsPOST /v1/shell-history for parsed Bash, Zsh extended-history, and Atuin records
These records use the same storage and correlation model as the rest of Cortex, which makes an agent change or shell command visible beside the service failure that followed it.
Fleet inventory
Inventory collection builds a redacted fleet snapshot from local files, SSH probes, Docker endpoints, and optional service APIs, then projects safe relationships into the investigation graph.
Investigation and intelligence
Cortex exposes bounded workflows rather than a raw SQL console.
Search and context
- FTS5 full-text search with host, app, severity, source, project, session, and time filters
- Structured filter-only retrieval for indexed fields
- Recent tails and single-row retrieval with raw-frame evidence
- Surrounding context around a log ID or timestamp
- Host, app, and source-IP inventories
- Clock-skew measurement using event and receive timestamps
Time and volume analysis
- Bucketed timelines
- Ingest-rate and queue-pressure state
- Near-duplicate message pattern clustering
- Recent-versus-baseline anomaly detection
- Side-by-side time-range comparison
- Silent-host and silent-stream detection
- Database, storage, and runtime statistics
Correlation
- Cross-host correlation around a timestamp
- AI-session anchor correlation against infrastructure logs
- Topic resolution through the entity graph before timeline construction
- Host-state correlation across logs, heartbeats, and inventory
- Historical incident similarity using FTS5
- Deterministic incident context bundles with bounded evidence
Recurring errors
An optional background scanner groups repeating error signatures into durable records. Operators can inspect unaddressed signatures, acknowledge them, revoke acknowledgements, and correlate a signature with logs and graph evidence.
Error detection is disabled by default. When enabled, it scans bounded batches, records lower-severity recurrences without paging, and can notify only above a configured severity floor.
Fleet inventory and graph
Inventory collectors
The native inventory subsystem can collect and normalize evidence from:
- Local and remote Compose, reverse-proxy, and AdGuard Home configuration
- Local process, storage, project, and raw configuration inventories
- SSH sessions to remote fleet hosts
- Local and remote Docker endpoints
- Tailscale
- UniFi
- Unraid
- Media-stack services and related APIs
SSH collection uses strict host-key verification, bounded concurrency, timeouts, and retry backoff. Sensitive fields are redacted before persistence.
The cache lives under ~/.cortex/inventory by default and includes:
normalized/homelab.json: the typed normalized fleet snapshotcollection-state.json: collector health, timing, and warning stateraw/<run-id>/...: raw-but-redacted supporting artifacts
The map action reads the normalized cache. It does not trigger a collection run and does not return raw config bodies or credential-bearing URLs.
Derived investigation graph
The graph connects canonical entities such as:
- Hosts and source identities
- Logical services and concrete service instances
- Applications and containers
- Domains, routes, and endpoints
- AI projects and sessions
- Error signatures and operational findings
Relationships carry confidence, trust, reason codes, timestamps, and bounded evidence references. The graph supports entity resolution, neighborhoods, topology questions, evidence lookup, and explanation paths.
The graph is a rebuildable projection. Raw logs, heartbeats, inventory records, error signatures, and AI session data remain authoritative. Projection rebuilds use staging tables and a short serialized swap, record watermarks and metrics, and preserve explicit degraded state when refresh fails.
AI session intelligence
Cortex treats AI transcripts as operational evidence, not merely chat archives.
Deterministic session analysis
The shared service layer can:
- List and search sessions by project
- Measure activity in five-hour usage blocks
- Summarize project context
- List observed tools and projects
- Detect frustration or abuse signals
- Group those signals into scored incidents
- Correlate AI activity with non-AI infrastructure logs
- Extract skill invocations
- Extract MCP server and tool-call events
- Extract hook configuration and runtime events
- Build skill-first, MCP-first, and hook-first investigation bundles
The deterministic query and incident workflows are available through CLI, REST, and MCP.
Guarded local assessments
LLM-backed assessments are deliberately local-only. They run through cortex assess and are not exposed as MCP actions or REST routes because they spawn a local provider subprocess.
CORTEX_LLM selects the provider and optional model for every LLM operation: skill, MCP, hook, and abuse assessments, including dry-run audit metadata. Set CORTEX_LLM=codex (the default, using Codex app-server) or CORTEX_LLM=gemini (Gemini CLI). To pin the model globally, use CORTEX_LLM=codex/gpt-5.5 or CORTEX_LLM=gemini/gemini-3.1-flash-lite-preview. An explicit model in CORTEX_LLM wins over --model; otherwise that existing flag can select a model for one invocation. Empty or unsupported selectors fail without provider fallback.
The selector also works in managed ~/.cortex/.env; process environment values take precedence. Settings are retained after config loading.
The old CORTEX_CODEX_MODEL and CORTEX_HEADLESS_GEMINI_MODEL variables are ignored. Command and authentication-home overrides remain installation settings, not provider selectors. --no-llm needs no provider configuration or credentials.
The shared LLM runner enforces:
- A global kill switch
- Global and per-action concurrency limits
- Per-minute and per-hour rate limits
- Per-action circuit breakers and cooldowns
- Invocation timeouts
- Prompt and output byte caps
- An explicit background-enrichment gate, disabled by default
- Durable audit records for successes, failures, timeouts, and policy denials
Default guard values allow one concurrent invocation, three per minute and thirty per hour per action, a 120-second timeout, a 1 MiB prompt cap, and a 256 KiB output cap.
Prompt scrubbing is enabled by default. Skill, MCP, and hook event extraction happens before scrubbed transcript text is persisted, so structured operational signals are retained without requiring raw prompt storage.
Alerts and notifications
Notifications are optional and disabled by default. When enabled, Cortex uses Apprise as the delivery bridge and a durable SQLite outbox for retry, deduplication, and dead-letter handling.
Built-in evaluators cover:
- OOM kills
- Containers exiting nonzero
- fail2ban bans
- Authelia MFA failures
- Disk-fill and storage guardrail pressure
- Ingest queue pressure
- Complete ingest silence
- Heartbeat silence
- Silence from previously active continuous streams
The notification subsystem includes:
- Configurable evaluator cadence
- Per-rule toggles and thresholds
- Deduplication windows
- Outage-scoped silence keys
- Bounded retry with dead-letter state
- Recent-firing history
- Test notifications
- A scheduled daily digest
By default, continuous stream-silence tracking covers UDP/TCP syslog, agent Docker, Docker stream and event records, and managed file tails. Sporadic sources such as transcripts and shell history are intentionally excluded.
Interfaces
CLI
Run cortex --help and command-specific --help for the generated command tree.
| Group | Purpose |
|---|
search, filter, tail | Log retrieval |
hosts, apps, entity, graph | Discovery and topology |
analysis, correlate, state, stats, timeline | Investigation and analytics |
sessions, assess | AI-session queries and local guarded assessments |
alerts | Error signatures, acknowledgements, and notification history |
ingest, heartbeat | Collectors, agents, file tails, inventory, and heartbeats |
serve, mcp | Full daemon and query-only stdio MCP modes |
doctor, status, db, compose | Diagnostics and maintenance |
setup, update, config, completions | Lifecycle and operator tooling |
Examples:
cortex search "oom killer" --host devhost --since 1h
cortex filter --severity err --since 6h
cortex timeline --since 24h
cortex sessions search "migration failure"
cortex graph explain --help
cortex alerts errors list
cortex ingest inventory refresh --json
cortex ingest filetail list
cortex status
cortex doctor
The CLI supports direct/local operation and HTTP operation. REST-backed mode is the normal remote path and uses CORTEX_URL plus CORTEX_API_TOKEN.
MCP
Cortex exposes one MCP tool named cortex. Its required action field selects an action from a single authoritative Rust registry. The mechanically generated current count is published in the live coverage inventory.
The current scope split is:
- 53 read actions requiring
cortex:read
- 6 admin actions requiring
cortex:admin: artifact_evidence_record, ack_error, unack_error, file_tails, notifications_test, and llm_invocations
- 1 informational action,
help, which requires an authenticated context when authentication is mounted but no read/admin scope
Complete MCP action catalog
| Domain | Actions |
|---|
| Log retrieval | search, filter, tail, errors, get, context |
| Discovery and health | hosts, apps, source_ips, status, stats, ingest_rate, silent_hosts, clock_skew |
| Analytics and correlation | timeline, patterns, anomalies, compare, correlate, topic_correlate, similar_incidents, recurring_error_comparison, incident_context |
| Fleet and topology | map, host_state, fleet_state, correlate_state, graph, compose_status, compose_doctor |
| AI sessions and scoped evidence | sessions, search_sessions, session_investigate, evidence_scope, abuse, abuse_incidents, abuse_investigate, ai_correlate, usage_blocks, project_context, list_ai_tools, list_ai_projects |
| AI operational events | skill_events, skill_incidents, skill_investigate, mcp_events, mcp_incidents, mcp_investigate, hook_events, hook_incidents, hook_investigate |
| Errors and administration | unaddressed_errors, ack_error, unack_error, notifications_recent, notifications_test, file_tails, llm_invocations, artifact_evidence_record |
| Reference | help |
The runtime schema contains per-action flags, defaults, examples, relative cost metadata, and validation. See docs/mcp/SCHEMA.md for the parameter reference.
MCP prompts
Cortex ships twelve reusable infrastructure prompts:
infra.incident-triageinfra.host-healthinfra.service-outageinfra.security-auth-reviewinfra.noise-reductioninfra.agent-change-correlationinfra.docker-container-regressioninfra.network-dns-failureinfra.storage-pressureinfra.auth-bruteforceinfra.syslog-forwarding-gapinfra.after-deploy-check
See docs/mcp/PROMPTS.md for arguments and output expectations.
MCP resources and UI
The MCP server exposes:
| URI | Purpose |
|---|
cortex://schema/mcp-tool | Live JSON schema for the action-dispatched tool |
cortex://schema/prompt-output | Schema for structured incident-style prompt output |
ui://cortex/query-widget | Self-contained MCP Apps search widget |
The query widget is progressive enhancement. UI-capable MCP hosts can render it; ordinary MCP clients continue receiving the normal text and structured JSON result.
REST
Authenticated JSON routes live under /api/* on the shared HTTP listener. They cover the same major query domains as MCP and add operator workflows for session checkpoints, parse errors, database integrity jobs, backup, checkpoint, and vacuum.
The versioned investigation API lives under /api/v1/* and provides Ask Cortex plus graph entity, neighborhood, explanation, and evidence endpoints for the bundled browser workspace.
REST requires CORTEX_API_TOKEN. Privileged maintenance and file-tail workflows can also require CORTEX_API_ADMIN_TOKEN.
See docs/api.md for the route and response reference.
Browser investigation workspace
The daemon serves a bundled workspace at /app and /app/investigate. It includes:
- Runtime and schema status
- Fleet and ingest summaries
- Ask Cortex investigation requests
- An interactive Cytoscape graph canvas
- Evidence inspection
- A recent-log timeline
The app is embedded into the Rust binary, has no external runtime dependency, uses a restrictive Content Security Policy, and keeps the entered REST bearer token in memory only. The current UI identifies itself as an investigation preview rather than a full general-purpose dashboard.
Claude Code plugin
.claude-plugin/plugin.json packages Cortex as a Claude Code plugin. It declares exactly three surfaces plus configuration:
| Key | Points at |
|---|
mcpServers | plugins/cortex/mcp.json — HTTP transport to ${user_config.server_url}/mcp with a Bearer ${user_config.api_token} header |
skills | plugins/cortex/skills/ — twelve skills |
userConfig | Server-vs-client mode, server_url, api_token, auth mode and OAuth fields, syslog/MCP bind host and port, retention and storage budgets, Docker ingest, and fleet hosts |
The twelve skills are cortex, frustration-assessment, hook-friction-assessment, incidents, logs, mcp-friction-assessment, report, searching-sessions, skill-improvement-assessment, topology, troubleshoot, and version-check.
The plugin registers no Claude Code lifecycle hooks — there is no hooks key and no hooks.json. Setup is explicit: run cortex setup pluginhook (or the plugins/cortex/scripts/plugin-setup.sh adapter) after installing, upgrading, or reconfiguring. just validate-plugin and scripts/validate-marketplace.sh assert the hooks key stays absent, and cargo xtask check-version-sync asserts the manifest carries no top-level version.
See docs/plugin/HOOKS.md for the setup lifecycle and docs/plugin/PLUGINS.md for the manifest reference.
Health endpoints
GET /health: minimal unauthenticated liveness response for containers and proxiesGET /health/full: detailed health and ingest observability; mounted auth requires CORTEX_TOKEN (OAuth alone is insufficient). Explicit loopback and trusted-gateway policies bypass this check.
Configuration
macOS heartbeat agent quick start
From an active desktop login:
cortex setup heartbeatagent install
cortex setup heartbeatagent check
Set CORTEX_HEARTBEAT_TARGET and an ingest credential first.
CORTEX_API_TOKEN is query-only and cannot authenticate agent ingest. See the
authoritative macOS heartbeat-agent operator
contract for capabilities, paths,
login semantics, migration, recovery, removal, and delivery proof.
Cortex loads configuration in this order, with later layers winning:
- Built-in defaults
- A partial
config.toml in the current working directory
- The managed
$CORTEX_HOME/.env file, normally ~/.cortex/.env
- Process environment variables
Important defaults
| Setting | Default |
|---|
| Syslog bind | 0.0.0.0:1514 UDP and TCP |
| HTTP bind | 127.0.0.1:3100 |
| Database path | /data/cortex.db unless setup or environment overrides it |
| SQLite pool size | 8 connections |
| SQLite page cache budget | 128 MiB total |
| SQLite mmap target | 256 MiB |
| Heavy-read concurrency | 1 |
| WAL checkpoint threshold | 256 MiB |
| Log retention | 90 days |
| Logical database limit | 1,024 MiB |
| Recovery threshold | 900 MiB, auto-adjusted when the limit is raised substantially |
| Cleanup cadence | 60 seconds |
| Notifications | Disabled |
| Recurring-error scanner | Disabled |
| Central Docker pull | Disabled |
| Prompt scrubbing | Enabled |
| Background LLM enrichment | Disabled |
Useful environment variables include:
| Variable | Purpose |
|---|
CORTEX_HOME | Managed configuration and operational home |
CORTEX_DB_PATH | SQLite database path |
CORTEX_RECEIVER_HOST, CORTEX_RECEIVER_PORT | Syslog listener |
CORTEX_ALLOWED_SOURCE_CIDRS | Optional syslog sender allowlist |
CORTEX_HOST, CORTEX_PORT | Shared HTTP listener |
CORTEX_TOKEN | Static MCP, OTLP, heartbeat, and forwarding bearer token |
CORTEX_API_TOKEN | REST bearer token |
CORTEX_API_ADMIN_TOKEN | Additional REST admin token where required |
CORTEX_AUTH_MODE | Static-token or OAuth authentication mode |
CORTEX_PUBLIC_URL | Public URL used for OAuth and exposure validation |
CORTEX_ALLOWED_HOSTS, CORTEX_ALLOWED_ORIGINS | Host-header and CORS policy |
CORTEX_RETENTION_DAYS | Log retention |
CORTEX_MAX_DB_SIZE_MB | Logical database storage budget |
CORTEX_MIN_FREE_DISK_MB | Optional external free-space write guard |
CORTEX_FILE_TAIL_ALLOWED_ROOTS | Managed file-tail root allowlist |
CORTEX_DOCKER_INGEST_ENABLED | Enable central Docker pull compatibility mode |
CORTEX_SCRUB_PROMPTS | Best-effort AI prompt credential scrubbing |
CORTEX_NOTIFICATIONS_ENABLED | Enable notification services |
CORTEX_LLM | Shared provider/model selector: codex[/MODEL] or gemini[/MODEL]; default codex |
CORTEX_LLM_ENABLED | Global local-assessment kill switch |
RUST_LOG | Tracing filter |
See docs/CONFIG.md for the complete reference and validation rules.
Authentication and trust boundaries
Cortex intentionally separates transport credentials and capabilities.
MCP
CORTEX_TOKEN authenticates static-token HTTP MCP calls.- Static-token callers receive
cortex:read by default.
- Set
CORTEX_STATIC_TOKEN_ADMIN=true only when the token should also receive cortex:admin.
- OAuth mode supports explicit read and admin scopes through the shared authentication layer.
- Non-loopback unauthenticated binds are rejected unless an explicit trusted-gateway mode is configured.
REST
CORTEX_API_TOKEN is separate from CORTEX_TOKEN.- REST refuses to mount without an API token.
- Sensitive maintenance and file-tail operations can require
CORTEX_API_ADMIN_TOKEN in addition to the normal REST bearer token.
Ingest endpoints
- Syslog itself is unauthenticated. Use CIDR and network controls.
- OTLP logs, metrics, and traces, plus heartbeats, AI transcripts, shell history, and agent-command forwarding, use the Cortex bearer-token policy when configured.
- Unauthenticated forwarding endpoints are loopback-only unless the operator explicitly establishes another trusted boundary.
Data handling
- AI prompt scrubbing is enabled by default.
- Inventory is redacted before it reaches the normalized cache.
- MCP and REST return bounded evidence instead of raw credential-bearing config artifacts.
- File tails are constrained to configured roots and reject path escapes.
- Docker endpoints and SSH keys are treated as privileged infrastructure access.
- Container examples mount a dedicated least-privilege SSH directory rather than a user's complete
~/.ssh.
Read docs/SECURITY.md, docs/GUARDRAILS.md, and docs/OAUTH.md before exposing Cortex beyond loopback or a trusted gateway.
Storage and maintenance
SQLite model
Cortex uses SQLite with:
- WAL mode
- A bounded r2d2 connection pool
- FTS5 external-content indexing with
logs_fts for the complete log corpus and a transcript-only ai_logs_fts projection that indexes message text plus provider scope for AI session search
- Covering and composite indexes for common filters and timelines
- Transactional batch writes
- Durable source checkpoints and parse errors
- Maintenance job tracking
- Online backup support
- Integrity checks, checkpoints, and vacuum workflows
The current schema history contains 63 sequential migrations. CI derives this denominator from KNOWN_SCHEMA_VERSION and the migration registry. Migration 63 adds indexed device attribution and a bounded, resumable background history pass that preserves raw authentication and session identity. Migration 60 introduced the transcript-only ai_logs_fts projection. Migration 61 rebuilds that derived index with provider scope and adds an AI-only (timestamp, ai_tool) index so time-bounded session searches can derive a safe rowid floor before FTS walks common terms. On large transcript histories these one-time derived-index rebuilds can hold the startup write transaction while they populate. Subsequent transcript inserts and deletes maintain ai_logs_fts incrementally, while global logs_fts remains the full-corpus search index. Server forwarding receipts have a seven-day replay horizon and are removed when their canonical evidence is deleted. The sender spool is intentionally shorter and bounded: an individual source retains at most 1,024 records or 1 MiB and evicts records older than one day; the aggregate spool retains at most 4,096 records or 4 MiB. Eviction removes the original payload and retains a pending gap marker so evidence loss remains visible. A retry after the server receipt horizon is a new ingestion attempt.
Authoritative and derived data
Authoritative records include:
- Normalized logs and raw frames
- Hosts and source identities
- Transcript source, checkpoint, import, and parse-error records
- Heartbeats and component snapshots
- Inventory snapshots and collector state
- Error signatures and acknowledgement events
- Skill, MCP, and hook events
- Notification outbox and firing history
- LLM invocation audit records
Derived accelerators include:
- AI session rollups
- Hourly timeline rollups
- Inventory statistics
- Stream-last-seen state
- Entity graph tables and evidence relationships
Derived data can be refreshed or rebuilt from authoritative evidence.
Storage guardrails
The daemon enforces retention and a logical database budget in chunks. It can enter a write-block state before uncontrolled growth damages the host, then resume after recovery thresholds are met. An optional minimum-free-disk guard protects against pressure caused by data outside the Cortex database without deleting Cortex data merely because another application filled the filesystem.
Operational commands:
cortex db status
cortex db integrity
cortex db checkpoint
cortex db backup --help
cortex db vacuum --help
Maintenance operations, including synchronous and background integrity checks, share one process-wide single-flight gate and are separately limited from heavy read queries. Concurrent attempts return a retryable busy response.
Deployment and distribution
Current identities
| Artifact | Current identity |
|---|
| Canonical source repository | dinglebear-ai/cortex |
| Native binary and CLI | cortex |
| npm launcher | @dinglebear/cortex |
| MCP Registry server name | ai.dinglebear/cortex |
| Published OCI image | ghcr.io/dinglebear-ai/cortex:v<version> |
The source repository and its published artifacts both live under the dinglebear-ai organization. The legacy jmagar namespace is retired: scripts/check-public-identity.sh fails the build on any tracked file that reintroduces it, and the release workflow derives the MCP Registry OCI identifier from the same REGISTRY/IMAGE_NAME it pushes to, so the two cannot drift apart again.
Container images published before the move remain readable under the legacy namespace for older pinned deployments, but nothing new is pushed there.
Native and npm
- The repository ships Linux x86_64, Linux aarch64, macOS arm64, and Windows x86_64 release assets.
- The npm launcher supports
linux/x64, linux/arm64, darwin/arm64, and win32/x64 and verifies release checksums.
- Source builds use Rust edition 2024 and the current stable toolchain in CI.
Docker Compose
The repository includes:
docker-compose.yml for a local build from the checkoutdocker-compose.prod.yml for the published imageconfig/Dockerfile for a non-root Debian runtime image
Manual Compose deployment expects an external Docker network named cortex unless DOCKER_NETWORK overrides it:
docker network inspect cortex >/dev/null 2>&1 || docker network create cortex
cp .env.example .env
bash scripts/prepare-compose-dirs.sh
docker compose up -d
curl -fsS http://127.0.0.1:3100/health
The preflight resolves the /backups bind with Docker Compose's own parser and
creates the default or CORTEX_BACKUP_DIR override at mode 0700. Compose is
configured not to create this host path implicitly, preventing root-owned or
overly permissive backup directories.
The Compose files:
- Publish syslog on UDP/TCP 1514
- Publish HTTP to host loopback by default
- Persist the database in a stable named volume or configured bind mount
- Run as a non-root UID/GID
- Mount inventory SSH credentials and workspace data read-only
- Mount a dedicated file-tail root read-only
- Use a 2 GiB default memory limit, configurable per host
- Include a health check and bounded container logs
For managed installation and repair:
cortex setup check
cortex setup repair
Host agents
Cortex includes setup and runtime support for satellite collection, including:
- Heartbeat agents
- Docker stream agents
- AI-session watch services and timers
- Shell-history forwarding
- Agent-command forwarding
- Completion and debug wrappers
Use cortex setup --help, cortex ingest --help, and cortex heartbeat --help for the exact platform-specific command tree.
Operations
Common operator commands:
cortex status
cortex doctor
cortex doctor binary
cortex setup check
cortex setup repair
cortex db status
cortex db integrity
cortex db checkpoint
cortex db backup --help
cortex db vacuum --help
cortex compose status
cortex compose doctor
cortex compose pull
cortex compose up
cortex compose restart
cortex compose logs
cortex ingest inventory status --json
cortex ingest inventory refresh --json
cortex ingest filetail list
cortex sessions doctor
cortex update --help
cortex config list
cortex completions zsh
The daemon exposes minimal and full health responses, records database maintenance jobs, reports projection and collector degradation, and surfaces ingest queue and write-block state through CLI, REST, and MCP.
Development and verification
The repository uses:
- Rust edition 2024
mise for pinned development toolsjust for common workflowscargo-nextest for the hermetic Rust suitecargo-llvm-cov for coveragecargo-deny for dependency policy- Lefthook and
cargo xtask for local release and pre-push checks
- A custom soldr-backed Rust compiler wrapper for fast local and CI builds
Common commands
mise install
just dev
just build
just check
just lint
just fmt
just test
just test-doc
just coverage
just coverage-html
just test-live
just validate-plugin
cargo xtask pre-push
just test-live (also just live-smoke) is the canonical fail-closed pull-request subset. It exercises real HTTP JSON-RPC, UDP and TCP syslog ingest, CLI/REST behavior, browser routes, and managed file-tail behavior in a run-owned topology. Run just live-mcp for every registered MCP action; the scheduled aggregate combines all authoritative owner profiles. Specialist profiles are documented in the live qualification guide. Docker collection has separate agent-deployment tests and a mocked Docker HTTP fixture for central pull.
CI gates include:
- Formatting
- Clippy with warnings denied
- Nextest and doctests
- Version and distribution identity synchronization
- MCP integration
- npm launcher checks
- Secret scanning
cargo-deny- Coverage generation
- Repository module-size policy
See tests/TEST_COVERAGE.md and docs/RELEASE.md for the split between hermetic CI and live-fleet verification.
Documentation
The code-owned registries and runtime schemas are authoritative for command names, actions, routes, scopes, defaults, and validation. Human documentation explains how to operate those surfaces.
Dated design plans and session logs preserve engineering history, not current public-interface authority. Runbooks are operational guidance unless explicitly marked historical; verify them against the current source and configuration.
Current boundaries
Cortex is intentionally opinionated:
- It is a single-node SQLite service, not a distributed ingestion cluster.
- OTLP logs, metrics, and traces are supported over HTTP/protobuf. OTLP/gRPC is outside the current implementation.
- Syslog does not authenticate senders. Network and CIDR controls matter.
- The graph is derived evidence, not authoritative configuration state.
- MCP and REST expose bounded operations, not arbitrary SQL or unaudited log mutation.
- LLM-backed assessments are local-only and require an operator-controlled installation of the provider selected by
CORTEX_LLM.
- Central Docker pull requires privileged read access to Docker endpoints and is disabled by default.
- Inventory quality depends on collector access, SSH trust, optional API credentials, and cache freshness.
- The bundled browser app is an investigation workspace preview, not a full monitoring dashboard.
- Cortex is built for a homelab or small trusted fleet. Internet-facing or multi-tenant deployment requires additional isolation and policy outside the binary.
License
Original Dinglebear-authored portions of this project are licensed under AGPL-3.0-only. Separate commercial licensing is available for organizations that need terms outside the AGPL. Third-party material remains under its original license. See LICENSING.md.