TECHNICAL CAPABILITY REFERENCE
Search the product capability guide across AI, discovery, detection, collection, access and operations. Use feature IDs to make technical evaluation precise.
AI-001
Model gateway (swap providers without rewiring) Select models and providers through a centralized gateway.
AI & Agentic
AI-002
Anthropic Claude routing (Sonnet, Haiku, Opus)
Access to Anthropic models.
AI & Agentic
AI-003
Google Gemini routing (Flash, Pro)
Access to Google models.
AI & Agentic
AI-004
OpenAI + Azure OpenAI routing Access to GPT models, incl. the enterprise Azure path.
AI & Agentic
AI-005
Azure AI Foundry (DeepSeek) routing Access to a lower-cost reasoning model.
AI & Agentic
AI-006
Local model routing (Ollama) Route workloads to local models through Ollama.
AI & Agentic
AI-007
Per-workload model selection Match model selection to the cost and reasoning requirements of each workload.
AI & Agentic
AI-008
Per-model rate and token limits Protects against runaway spend and provider throttling.
AI & Agentic
AI-009
Input-size guard (200k token cap)
Stops an accidental log dump from becoming a huge bill.
AI & Agentic
AI-010
Cross-provider parameter compatibility
One code path works across all providers.
AI & Agentic
AI-011
Structured / schema-enforced model output
Require structured model responses that conform to the expected output schema.
AI & Agentic
AI-012
Native tool-calling support
Invoke platform functions through native model tool calling.
AI & Agentic
AI-013
Per-request token accounting Track token usage for each AI request.
AI & Agentic
AI-014
Cumulative token tracking per investigation Track cumulative token consumption across an investigation.
AI & Agentic
AI-015
Live token counter exposed as a status query Watch spend while an investigation runs.
AI & Agentic
AI-016
Per-virtual-key dollar budgets Apply spend caps by customer or team.
AI & Agentic
AI-017
Embeddings / vector search / RAG retrieval
Semantic search over the customer's own data.
AI & Agentic
AI-018
MCP tool-server support Standardised way to plug external tools into the agent.
AI & Agentic
AI-019
Automatic conversation compaction
Summarize conversation history to keep long investigations within model context limits.
AI & Agentic
AI-020
LLM-written running summary of an investigation
Preserves findings, hypotheses and data points while dropping bulk.
AI & Agentic
AI-021
Per-step context windowing Each reasoning step only sees what it needs.
AI & Agentic
AI-022
Tool-result truncation Bound tool-response size to protect conversation context.
AI & Agentic
AI-023
Token estimation before calling the model Estimate token requirements before submitting a model request.
AI & Agentic
AI-024
Tool: search_events The agent can search the customer's events by keyword, time, severity and source.
AI & Agentic
AI-025
Tool: get_metrics
The agent can pull time-series data for a device or interface.
AI & Agentic
AI-026
Tool: get_top_talkers
The agent can find the busiest source IPs.
AI & Agentic
AI-027
Tool: aggregate_metrics The agent can compute averages, maxima, minima and sums itself.
AI & Agentic
AI-028
Tool: search_logs Full-text search across all log content.
AI & Agentic
AI-029
Tool: get_device_summary Events, alert counts and metric baselines for one device in one call.
AI & Agentic
AI-030
Tool: ip_lookup (geo + ASN) The agent can find out where an address is and who owns it.
AI & Agentic
A few good questions
Get clear
on the details.
Can we import every spreadsheet or portal automatically?
Collection and onboarding depend on the source. Validate each connection and mapping; do not assume a universal spreadsheet or portal importer. Operational knowledge can be added through the supported context and operator-guidance workflows.
Your next operational advantage
Put AI to work.
Leave the sprawl behind.
Bring your stack and a recent incident.
See what your team can simplify.