Compose
Pick an agent template (50+ available), customize identity + tools, attach a runtime policy.
Solution: Agent platform
Ship AI agents to production without building governance from scratch.
GoodHelp is the multi-agent platform with action ledger, SoD approvals, runtime policy, and signed evidence already in production. You bring the use case; we bring the controls.
50+ agent templatesHash-chained action ledgerSigned evidence pack
When you have AI agents doing real work, you want every action logged, gated, and exportable, so that your compliance officer doesn't kill the project the first time someone asks 'what did the agent do?'
How it works
From template to production in three steps.
Approve
Agents request approval at the SoD-policy-defined gates. HITL by default. Enforced fail-CLOSED (SOD_REQUIRE_AUDIT_AVAILABLE=true): if the SoD audit context is unavailable the approval is blocked rather than allowed.
Audit
Every action enters the hash-chained ledger. Signed evidence pack exports on demand — JSON manifest plus body, KMS-rooted signing key.
Capabilities
The control envelope you would otherwise build yourself.
Multi-agent orchestration
Sub-agents compose into department agents (CMO, CFO, CRO, CTO, COO). Deterministic composition, not free-form delegation.
Action ledger
Hash-chained, KMS-rooted, exportable as a signed evidence pack — JSON manifest plus body (CSV/TSV/JSON).
Runtime policy
Per-agent constraints on tools, external-comms allow-list, and per-run cost cap. Enforced server-side.
Connectors
Pre-built integrations across CMS, social, billing, observability, and LLM providers — see the integrations catalog for the current surface.
Use cases
Where teams put the platform to work.
Operations
Dunning, vendor management, vendor invoice triage — all under SoD approval gates.
Marketing
CMO + Content Autopilot run the editorial calendar across every venture on a weekly cadence.
Finance
Monthly close helper, cash position dashboard, Mercury read-only balance + transaction feed.
Engineering
DORA metrics, CI monitoring, code review triage — composed CTO agent with focused sub-agents.
Sales
Pipeline scoring, lead routing, churn-risk surfacing — material actions never bypass an approver.
Already in production.
17
Production agents running LMNTL itself
50+
Agent templates in the seed library
8 engineer-days
LMNTL's COO playbook transition to agents
Where GoodHelp sits relative to your other options.
| Capability | GoodHelp | Generic agent SDKs (LangGraph, AutoGen, CrewAI) | Zapier AI / Lindy |
|---|---|---|---|
| Deterministic workflows (Temporal-backed) | ✅ | DIY | ❌ |
| Hash-chained action ledger | ✅ | DIY | ❌ |
| SoD approvals | ✅ | DIY | ❌ |
| Signed evidence export | ✅ | DIY | ❌ |
| Built for HITL | ✅ | Varies | ❌ |
Comparison reflects publicly documented capabilities as of 2026-05-25. Customer mileage may vary by configuration.
What teams ask first.
Is this a framework or a platform?
Platform. You configure agents; we run the workflows on Temporal-backed deterministic infrastructure with the action ledger and approval gates already wired.
Do you support custom agents?
Yes — via the agent builder. Custom skills require a platform-team conversation today.
What LLMs do you use?
Multi-provider (OpenAI, Anthropic, Google). See the platform agents page for model routing.
Can I bring my own LLM?
Not today. On the roadmap.
Pricing?
Per-agent + LLM passthrough. See the pricing page.
Run agents your auditor can follow.
Thirty minutes with a Trust Center walkthrough — bring your governance checklist; we’ll walk through the controls one by one.