Scale Your Operations
With Autonomous Agents.
Deploy 250+ AI agents across 8 business departments, then govern how they execute inside your systems with approvals, policies, and tool access.
Your teams have AI tools.
The work still backs up.
Most companies bought chat interfaces, not accountable execution. Sentrul gives service, revenue, IT, and operations leaders a governed AI workforce that can take action inside the systems they already run.
Point solutions that stop at answers instead of completing work
Automation without the approvals, audit trails, and controls operators need
Sentrul: 250+ AI specialists across 8 business departments
From Operating Context to Live Work in 3 Steps.
Connect tools, choose the right department launch team, and put governed agents into production fast.
Connect & Inform
Connect your systems with MCP, add operating context, and ground each workflow in the way your team already works.
Launch the Right Team
Start from 254 featured specialists across 8 business departments, then tailor instructions, approvals, and tool access for each team.
Run with Oversight
Put agents into live queues while keeping high-stakes actions behind human review, audit trails, and policy controls.
Research Copilot
AI-Powered Research, Grounded in Your Data
Hybrid search combines vector embeddings with keyword matching. Your agents cite sources, not hallucinations.
Hybrid Search
pgvector + Algolia: semantic and keyword search combined for maximum recall.
Semantic Memory
Embeddings stored per tenant. Your agents remember context across sessions.
Document OCR
Upload PDFs, images, and scanned docs. Text extraction is automatic.
Source Citations
Every answer links back to the original document chunk it drew from.
Enterprise Security
Enterprise-Grade Security, Built In
Row-level isolation, tenant-scoped encryption, and a fail-closed safety pipeline. Your data never crosses tenant boundaries.
AES-256 Encryption
Data encrypted at rest and in transit. Azure Key Vault manages all secrets.
3-Tier Guardrails
NeMo Rails input filter, Azure Content Safety, and output policy checks run on every request.
PII Redaction
Automatic detection and masking of personal data before it reaches any model.
Bring Your Own Key
Use your own API keys for LLM providers. Your keys, your control, your spend.
Human Oversight
Choose Oversight Mode by Department
Pick how much autonomy each team gets in production, then tune that mode by workflow as confidence grows.
Human-in-the-Loop
Every agent action requires explicit human approval before execution.
Human-on-the-Loop
Agents execute autonomously but pause on high-confidence-threshold decisions for review.
Human-out-of-the-Loop
Full autonomous operation with post-execution audit trail and emergency stop.
Production Reliability Without Guesswork.
Combine the oversight modes above with deterministic routing, escalation controls, and verifiable execution logs.
Deterministic Failover
If a model or lane degrades, routing falls back through healthy paths so critical work keeps moving.
Confidence-Gated Escalation
Low-confidence outputs are automatically surfaced for review so your team can intervene before downstream impact.
Signed Audit Trail
Execution records include integrity hashes and signatures so compliance teams can verify exactly what happened.
250+ AI Agents Across 8 Business Departments.
Launch customer service, revenue, IT, marketing, HR, supply chain, and construction workflows from one governed operating layer.
+ up to 3 custom agents per workspace (more on higher tiers)
Customer Service
42 Agents
Ticket triage, escalation coverage, and QA support
Sales & RevOps
55 Agents
Pipeline follow-up, proposal support, and CRM hygiene
IT Security
25 Agents
Alert triage, access review support, and evidence coordination
Marketing
27 Agents
Campaign operations, content workflows, and attribution support
IT Operations
24 Agents
Incident coordination, runbook execution, and service desk coverage
Human Resources
18 Agents
Hiring coordination, onboarding, and policy triage
Supply Chain
33 Agents
Exception handling, vendor coordination, and order reconciliation
Construction
30 Agents
Bid support, permit tracking, and field coordination
Counts reflect the live agent catalog. Some agents are available only on specific plans.
Professional Pricing for Every Scale.
Start free, upgrade when you need more capacity. No credit card required.
Trial
Explore the workspace with your own provider keys before moving onto managed capacity.
- 2 active agents, 1 member, 1 department
- 1 MCP connector
- BYOK required with 5M token cap per billing period
- Tier 1 local, Tier 2 mid-cost, and Tier 3 premium lanes
- No managed tokens or audit-log retention
- Up to 3 custom agents
Starter
First paid tier for small teams that want managed tokens without committing to enterprise controls.
- 3M managed tokens per month
- 2 active agents and 3 members
- 1 department and 3 MCP connectors
- 30-day audit logs
- BYOK fallback after managed exhaustion
- Up to 3 custom agents
Essentials
Department-level automation for teams that need document ingestion, broader collaboration, and higher throughput.
- 15M managed tokens per month
- 10 active agents and 10 members
- 3 departments and 12 MCP connectors
- Document uploads and 90-day audit logs
- Boost token-pack access and BYOK fallback
- Up to 6 custom agents
Operations
Most Popular
Cross-functional production tier with premium model access, compliance features, and parallel BYOK.
- 75M managed tokens per month
- 25 active agents and 25 members
- 10 departments and 24 MCP connectors
- Tier 1 local, Tier 2 mid-cost, and Tier 3 premium lanes
- Parallel BYOK, cost-aware routing, 365-day audit logs
- SSO, private VPC option, and HIPAA BAA availability
- Up to 12 custom agents
Enterprise
Top tier for organizations that need isolation, unlimited scale markers, and priority handling.
- 100M managed starter pool
- Unlimited agents, members, departments, and connectors
- BYOK-primary execution
- Permanent audit trails and priority command lanes
- Dedicated AKS isolation, private network options, and priority support
- Unlimited custom agents
Frequently Asked Questions
- What are managed tokens?
- Managed tokens are Sentrul-provided monthly AI budget for supported plans. They are separate from BYOK traffic and are consumed faster on higher-cost model lanes.
- How does BYOK work by plan?
- Trial requires BYOK. Starter and Essentials use managed tokens first and can fall back to BYOK. Operations can run managed and BYOK paths in parallel. Enterprise is BYOK-primary with negotiated overage handling.
- How do model lanes affect spend?
- Managed usage is weighted by lane: tier 1 is 1x, tier 2 is 3x, and tier 3 is 10x. Premium-heavy workloads consume the managed pool faster than low-cost routing.
- What happens near the end of my monthly budget?
- Routing can downgrade premium requests as managed utilization rises, and final exhaustion behavior depends on your plan and whether valid BYOK keys are available.
- Which plans can buy token packs?
- Token packs are gated by plan: Essentials can buy Boost, while Operations and Enterprise can buy Boost, Scale, and Surge packs.
- Why don't I see a per-model dollar estimate?
- Your cost on Sentrul is based on your plan's included tokens and a fixed multiplier per model lane — not the raw Azure Foundry wholesale rate. Showing wholesale prices would misrepresent your actual cost. See your per-lane multiplier on the Model Router tab for any agent.
- Do you claim SOC 2 certification?
- Public documentation should treat Sentrul as providing audit-ready controls and plan-dependent compliance features, not as a blanket certification claim.