This project is under active development and is not yet production-ready. APIs, table schemas, and class interfaces may change before the first stable release.
SNOllamaCore is a ServiceNow scoped application that integrates one or more self-hosted Ollama AI servers with the ServiceNow platform. It provides connection management, model lifecycle operations, session-aware AI inference, and a clean public API callable from any ServiceNow scope.
The goal is to make it straightforward to send prompts to locally-hosted LLMs from Business Rules, Flow Designer actions, Script Includes, and any other ServiceNow automation context — without exposing data to third-party cloud AI providers.
- Server Instance management — store, test, and manage connections to one or more Ollama servers, with optional MID Server routing and Bearer token authentication.
- Model sync — discover and synchronize available models from Ollama into ServiceNow records, with automatic context cards populated on sync.
- Prompt templates — define reusable, parameterized prompts with
${variable}substitution, a recommended model, and sampling settings. Execute them by SysID from anywhere on the platform. - Generate — single-turn text generation driven by a prompt template. Every call creates a session and persists the prompt and response for audit.
- Chat — multi-turn conversation with automatic session history. Resume an existing session to carry context across calls.
- Public API — a single class,
OllamaApi, is accessible from any scope. All other classes are package-private.
For a full walkthrough of the architecture, table schemas, class reference, configuration steps, and working code examples, see the implementation guide:
docs/SNOllamaCore_Implementation_Guide.md
| Scope | x_404031_snollam_0 |
| Version | 1.0.0 |
| Platform | ServiceNow Washington DC+ |
| JS Level | es_latest |