Core Dependencies

ComponentSupportedTested WithNotes
Python3.113.11Bundled in Docker image
PostgreSQL15, 1616pgvector extension required
Redis7.x7 (Alpine)Used for caching
Docker Engine24+27.xDocker Compose v2 required
Docker Composev2.20+v2.32v1 not supported

Container Orchestration

PlatformSupportedTested WithNotes
Docker Composev2.20+v2.32Dev / eval deployments
k3s1.28 – 1.321.32Production recommended
Kubernetes1.28+1.32 (k3s)Helm chart or Kustomize
Helm3.12+3.16Chart version 0.1.0

Digital.ai Release

ComponentSupported VersionsNotes
Release23.3, 24.1, 24.3, 25.1MCP server connects via REST API
Release MCP ServerBundledRuns as sidecar container

Jira (Optional)

ComponentSupportedNotes
Jira CloudCurrentAPI token authentication
Jira Data Center9.x+API token or basic auth
Jira MCP ServerBundledRuns as sidecar container

LLM / AI

Local Models (Ollama)

ComponentSupportedRecommendedNotes
Ollama0.5+LatestHosts local LLM models
Router modelllama3.2:3bllama3.2:3bFast intent classification
Agent modelqwen3:14bqwen3:14bTool-calling, format compliance
Embedding modelnomic-embed-textnomic-embed-textCross-session memory

Cloud LLM Alternatives

ProviderSupport LevelConfig Variable
Anthropic ClaudeSupportedREVA_ANTHROPIC_API_KEY
OpenAISupportedREVA_OPENAI_API_KEY
vLLMExperimentalREVA_VLLM_URL

When using a cloud LLM, the local GPU/Ollama requirement is removed. The router model (llama3.2:3b) still needs Ollama.

GPU Requirements

GPUVRAMSuitabilityNotes
NVIDIA RTX 5070 Ti 16GB16 GBMINIMUMSingle-user only, ~22s response time
NVIDIA RTX 5080 16GB16 GBRECOMMENDEDFaster inference, single-user or light concurrency
NVIDIA RTX 509032 GBRECOMMENDEDHeadroom for concurrent users
NVIDIA A1024 GBRECOMMENDEDData center, good throughput
NVIDIA L40S48 GBOPTIMALMulti-user, parallel inference viable

Important: Do NOT set OLLAMA_NUM_PARALLEL=2 on 16 GB GPUs — it causes 5× slowdown due to KV cache overflow.

Driver Requirements

ComponentMinimum Version
NVIDIA Driver535+
CUDA12.1+
NVIDIA Container ToolkitLatest

Microsoft Teams / Azure

ComponentSupportedNotes
Microsoft TeamsCurrent (desktop, web, mobile)Bot Framework v4
Azure Bot RegistrationSingle-tenant, multi-tenantREVA_MICROSOFT_APP_TYPE
Azure ADCurrentJWT validation for incoming activities

LDAP (Optional)

ComponentSupportedNotes
Active Directory2016+LDAP / LDAPS
OpenLDAP2.xFor group/role resolution

Operating System (Host)

OSTested WithNotes
Ubuntu22.04, 24.04Recommended for production
Debian12Supported
RHEL / Rocky8, 9Supported

Reva runs in containers, so the host OS matters only for Docker/k3s and NVIDIA driver support.

This website does not use cookies or tracking technologies. All fonts are self-hosted; no data is transferred to third parties. See our Privacy Policy for details.