Configuration

MAPLE configuration should be explicit enough that an operator can answer three questions quickly:

1. Which services are enabled?
2. Which models and policies are default?
3. Which deployment profile is active?

Example config

profile = "standard"
endpoint = "http://localhost:8080"
registry = "registry.mapleai.org"

[model]
default_backend = "ollama"
default_model = "llama3.2:8b-q4"

[guard]
default_policy = "policies/default.yaml"
pii_redaction = true

[fleet]
budget_monthly_usd = 2000
canary_enabled = true

Useful environment variables

| Variable | Meaning | | --- | --- | | PALM_ENDPOINT | CLI and SDK endpoint override | | MAPLE_PROFILE | Active deployment profile | | MAPLE_MODEL_BACKEND | Default routing backend | | MAPLE_MODEL_NAME | Default model identifier | | MAPLE_GUARD_POLICY | Guard policy override |

Deployment profiles

Minimal

Single-node development with lightweight persistence.

Standard

Shared team environment with observability and policy simulation.

Financial

Higher approval, audit, and idempotency requirements.

Sovereign

Private registry, mirrored artifacts, and controlled connectivity.

Federated

Multi-domain deployment with stronger tenancy boundaries and mirrored policy packs.