Compare statements
PDF Diff compares two versions of a financial statement, policy, or filing on your own computer and exports a highlighted PDF. The sample comparison uses two quarters of a public company’s statements. Free for Mac and Windows.
An AI policy needs somewhere to run.
Banks and credit unions are writing AI acceptable-use policies while chatbots multiply across CRM, contact center, and core systems. WolfPackAI gives that policy a technical home: one gateway on your infrastructure that decides which models are available and shows how they are used.
What your policy asks for, and where we are today.
AI policies tend to ask for the same few things. Here is how each one is supported, including the parts that are still on the roadmap.
| Your policy says | How it is supported |
|---|---|
| Only approved models | WolfPackAI routes model use through one gateway, so the models available to staff are set in one place. Live in the alpha today. |
| Sensitive data stays in-house | Open-weight models run on your hardware. Only the requests you choose go to commercial providers. Live in the alpha today. |
| We can see how AI is used | OpenTelemetry traces, metrics, and logs for each service today. Richer monitoring is next on the roadmap. |
| Access by role | Basic authentication today. Role-based access control is planned for the WolfPackAI beta. |
| A person approves consequential actions | Built into each engagement: anything that moves money or changes a record waits for a named person. Where people step in. |
It sits above your systems. Nothing gets ripped out.
The gateway sits between your people and applications and the models they call. CRM, contact center, and core systems keep running as they do today. WolfPackAI in detail
- Deploys on existing hardware, new servers, private cloud, or hybrid
- n8n workflows and webhooks into the systems you already run
- Local models when latency or data residency calls for them
- A path from a discovery pilot to hardened production
Current status
WolfPackAI is in alpha. Model routing, basic authentication, and tracing work today and are ready for a pilot. Role-based access control, richer monitoring, and Kubernetes deployment are next on the roadmap.
Teams that need an assistant of their own can add NoPilot, a Copilot-style assistant you host.
Statements, filings, and policies, handled locally.
Search policies and procedures
Private RAG answers questions from credit policies, procedure manuals, and filings, with sources, on infrastructure you control. How it works
Extract from filings
For our customer Ebullience, Palm Hill built a platform that extracts executive compensation from SEC filings into structured data. The same document work is available as an engagement. See the Ebullience work
Questions from banks and credit unions
Does WolfPackAI enforce our AI acceptable-use policy today?
Partly. Today it routes model use through one gateway on your infrastructure and traces each service, so approved models and their use are in one place. Role-based access control is planned for the beta release. We help you decide which parts of the policy the gateway should carry and which stay with people and process.
Do we have to replace our CRM, contact center, or core systems?
No. The gateway sits above them, and nothing gets ripped out. Those systems reach models through one route instead of many.
Can we still use commercial models?
Yes. Commercial models are reached through the same gateway as open-weight models on your hardware. Only the requests you choose to route to them leave your environment.
How do you handle SOX and GDPR?
We design with SOX and GDPR requirements in mind and review the architecture with your compliance team. We don’t certify compliance; we build systems your compliance team can review.
Where should we start?
With an inventory of where AI is already in use and a gateway pilot on one team’s workflows. If you want to see local processing first, PDF Diff is a free download.
Bring the policy. We will show where it can run.
We will map your AI policy to a gateway pilot on your infrastructure and be clear about what is live and what is roadmap. We don’t partner with any chip, hardware, or cloud vendor, so where it runs is your choice.