Features

Everything required for AI search that actually ships to production.

Findola is not a model; it's the layer around the model — index, permissions, agents, observability, evals — built so your team can trust every answer.

Core

Permission-preserving retrieval at scale.

ACLs from every source system are honored at query time, not index time. New permissions take effect within seconds — no full re-index, no leakage windows.

Used by every healthcare and financial-services customer in production today.

Core

Cross-app entity resolution.

Customers, deals, tickets, repos and people unified across systems with a single canonical ID.

Core

Cited answers.

Every claim links to its source. Confidence-scored. Findola abstains when retrieval is weak.

Agents

Planner · Executor · Verifier.

Multi-agent orchestration with bounded retries and escalation. Verifier model is independent of the executor to catch hallucinations before they reach the user.

Observability

Trace every step.

Per-request latency, cost, token usage, and tool calls — visible in the dashboard and exportable to your warehouse.

Eval

Golden datasets.

Bring your own evals; run on every prompt or model change before it ships.

Multi-modal

PDFs, screenshots, voice.

Findola searches the original layout, not just OCR text.

Admin

Policy engine.

Allowlists, blocklists, budget caps, and two-person approval for sensitive actions.

Feature index

A complete list of what ships today.

Categories below summarize the production feature set across the search, agent, admin, and analytics surfaces.

Search & retrieval
  • · Hybrid BM25 + dense + late-interaction reranker
  • · Per-document ACL enforcement
  • · Incremental indexing via CDC
  • · Multilingual embeddings
  • · Layout-aware document parsing
  • · Synonym & entity boosting
  • · Query-aware filtering & faceting
Agentic workflows
  • · Multi-agent planner / executor / verifier
  • · Drag-and-drop workflow builder (V2)
  • · Natural-language workflow authoring
  • · Trigger, schedule, and conversation-based
  • · Library of pre-built templates per vertical
  • · Human-in-the-loop review UI
  • · Long-term episodic memory per user & org
Admin & governance
  • · SSO / SAML / SCIM via WorkOS
  • · Role-based access control
  • · Immutable audit log (SIEM export)
  • · Connector management & OAuth scopes
  • · Policy editor (allow / block / escalate)
  • · Spend caps and budget alerts
  • · Per-tenant KMS keys (BYOK)
Analytics
  • · Per-workflow KPIs & ROI calculator
  • · Per-user activity dashboards
  • · Cohort adoption & stickiness
  • · Drill-down to any agent action
  • · Executive weekly digest (AI-generated)
  • · Export to Snowflake / BigQuery / S3
Developer surface
  • · REST + Webhook + MCP API
  • · TypeScript & Python SDKs
  • · Custom skill / tool registry
  • · Eval harness (CI integration)
  • · Sandboxed tool execution
  • · Public-facing changelog & status page
Enterprise
  • · Custom fine-tuned model per tenant
  • · On-prem / VPC / BYOC deployment
  • · Dedicated GPU pool, SLA-backed
  • · Forward-deployed engineering team
  • · Multi-region data residency
  • · Industry compliance packs (HIPAA, FedRAMP, PCI, GxP)

See it on your own data.

Connect three apps and Findola will answer your first question in under ten minutes.