FAQ
AdvisorClaw FAQ
A standalone FAQ covering the core product, deployment, architecture, privacy, and model-provider questions most firms ask first.
FAQ
AdvisorClaw is a dedicated AI hub built for advisory firms, deployed on a dedicated server for each client.
Rather than giving firms a generic chatbot inside a shared SaaS application, AdvisorClaw deploys a private, customized environment with Enhanced Memory, a live workspace, advisor-specific skills, and advisor-focused tools.
A chatbot answers prompts. A note taker records conversations. AdvisorClaw is designed to do significantly more.
Each deployment includes a persistent workspace, Enhanced Memory, tool-enabled workflows, specialist roles, and durable files and session context.
That means it can support real multi-step work across research, operations, portfolio analysis, and compliance-oriented workflows.
AdvisorClaw is deployed on a dedicated VPS for each client rather than inside a shared multi-tenant AI application.
That creates several advantages: isolated runtime per customer, stronger separation of files, memory, and runtime state, tighter control over model-provider access, cleaner operational boundaries for sensitive workflows, and less dependence on shared application infrastructure.
For firms that care about control, privacy posture, and deployment isolation, this is a materially stronger model than lightweight shared-server AI tools.
Yes. AdvisorClaw is deployed as a dedicated, single-instance environment for each client.
That means each customer receives a private operating environment rather than sharing a common runtime with other customers.
A standard AdvisorClaw deployment includes a controlled reverse-proxy edge, the customized AdvisorClaw application layer, the agent runtime and gateway, seeded workspace and memory structures, a main assistant, specialist roles for research, portfolio analysis, and compliance-oriented work, advisor-specific skills and tools, and deployment validation before handoff.
AdvisorClaw is built for advisor-facing workflows such as market and macro research, portfolio analysis support, compliance-oriented review workflows, recurring internal operations, document-heavy tasks, and structured knowledge work that benefits from persistent context.
The platform is especially useful where files, memory, and repeatable workflows matter.
In practical terms, Enhanced Memory uses a durable file-backed structure informed by QMD-oriented patterns.
That memory architecture is designed to support continuity across sessions, accumulation of firm-specific context, persistent operating knowledge, and structured retrieval of prior work.
It gives the deployment a durable working memory rather than forcing the system to start from zero every time.
Yes. Customers can upload and download files into and out of the workspace and session environment.
That is intentional. AdvisorClaw is designed to integrate with real workflows rather than trap work inside a closed interface.
Runtime data lives inside the deployed AdvisorClaw environment.
That includes workspace files, session artifacts, local memory structures, logs, and agent-specific workspaces.
This keeps the operating context tied to the dedicated deployment rather than collapsing everything into a centralized shared application layer.
Current deployment paths support direct integration with OpenAI and Anthropic.
AdvisorClaw is designed as an open architecture, so model-provider configuration can be shaped around customer requirements and deployment preferences.
No. AdvisorClaw is designed to be model-provider-flexible.
It is a customizable AI operating framework rather than a closed-model application. That gives customers more flexibility around provider selection, privacy posture, and workflow design.
AdvisorClaw inherits the privacy characteristics of the underlying model providers used through API access.
In practice, the deployment can be configured to call approved model vendors, giving customers more control than they would typically have inside a generic shared AI SaaS product.
AdvisorClaw is designed as a deployed runtime environment rather than a centralized prompt-routing SaaS layer.
The platform’s operating model emphasizes local workspace state, configurable provider access, and dedicated deployment boundaries rather than a shared central prompt-handling system.
The dedicated VPS model creates several important advantages: stronger runtime isolation, private workspace and session state, better separation from other customers, more controllable model-provider access, and easier customization for firm-specific workflows.
For advisory firms, this is a much better fit than generic AI products built for broad shared usage.
No. Each client receives a dedicated, single-instance environment.
That means your runtime, workspace, memory, and operating context are associated with your own deployment.
Current deployments use authenticated application access together with protected internal gateway configuration.
The architecture also includes a separated public edge, isolated internal runtime services, deploy-time secret injection, service-health and workspace-integrity validation, and seeded permission and ownership controls.
AdvisorClaw is designed for advisory workflows, where firms often want stronger architectural boundaries around AI use, more control over files and workflow state, better separation from shared AI app environments, support for document-heavy and knowledge-heavy work, and a deployment they can shape around real operating needs.
That makes it especially relevant for firms that want AI capability in a more controlled form.
Yes. Current deployments support a primary assistant plus specialist roles for market research, portfolio analysis, and compliance-oriented work.
This creates a more structured and useful operating model than a single generic assistant.
Yes. AdvisorClaw is designed for integration rather than lock-in.
It can work with model-provider APIs, file-based workflows, deployment/control-plane systems, customer-specific skills and tool layers, and enterprise processes built around documents, local runtime context, and durable work products.
The current implementation includes integration with DigitalOcean for deployment provisioning, Supabase for control-plane persistence, OpenAI and Anthropic for model-provider access, GitHub-hosted repositories for deployment-time assembly, and Resend for access-request notification workflows.
AdvisorClaw is currently delivered through a managed private deployment model.
That means deployments are provisioned intentionally, shaped for each client, and validated before handoff rather than launched through an instant self-serve flow.
At this stage, managed rollout allows tighter deployment quality, better security hygiene, stronger environment shaping, more tailored skills and workflow configuration, and better alignment with each firm’s actual use cases.
IDX is developing IDX Intellex, a self-hosted LLM offering intended to become an optional model path for AdvisorClaw deployments.
IDX Intellex is based on current open-source frontier models and is being customized for financial use cases.
IDX Intellex is expected to provide lower effective token cost, more processing capacity for high-volume workflows, better economics for research-heavy and document-heavy use, and stronger financial-domain alignment.
It should be understood as an upcoming capability rather than the current default deployment baseline.
The strongest reason is architectural.
AdvisorClaw gives firms a private, advisor-focused AI operating environment with dedicated infrastructure, Enhanced Memory, advisor-specific skills and tools, configurable model access, and stronger control boundaries than typical shared-server AI products.