OSS / Getting Started

Editions

What is available in Aurora OSS and Aurora Enterprise, and how to start with the right edition.

Browse docs

Aurora ships as two launch artifacts with the same OpenAI- and Anthropic-compatible gateway core and different capability gates.

Edition	Artifact	Best for	License
OSS	`aurora-oss` binary or public container	local teams, simple production gateways, provider aggregation, observability, and policy basics	not required
Enterprise	Provided by Aurora during onboarding	organizations needing identity, RBAC, tenant administration, budget enforcement, advanced routing, compliance, HA controls, and export connectors	license and access handled by Aurora

OSS features

These features are available without an Enterprise capability:

Feature	Configure	Use
OpenAI- and Anthropic-compatible gateway	provider env vars or `providers:` YAML	call `/v1/chat/completions`, `/v1/responses`, `/v1/embeddings`, `/v1/models`, `/v1/files`, `/v1/batches`, and `/v1/messages`
Provider auto-discovery	provider env vars such as `OPENAI_API_KEY`, `ANTHROPIC_API_KEY`, `GEMINI_API_KEY`, `OLLAMA_BASE_URL`, or `VLLM_BASE_URL`	list models at `/v1/models` and request them by model ID
Provider passthrough	`ENABLE_PASSTHROUGH_ROUTES=true` and `ENABLED_PASSTHROUGH_PROVIDERS`	call `/p/{provider}/...` for native provider APIs
Admin dashboard and API	`ADMIN_UI_ENABLED=true`, `ADMIN_ENDPOINTS_ENABLED=true`	open `/admin/dashboard` or call `/admin/api/v1/*` with the master key
Managed API keys	dashboard `Auth Keys` or `/admin/api/v1/auth-keys`	issue gateway keys without sharing upstream provider keys
Usage and cost analytics	`USAGE_ENABLED=true`	view `Usage` and `/admin/api/v1/usage/*`
Audit logs and live console	`LOGGING_ENABLED=true`	view `Audit Logs`, `Console`, and `/admin/api/v1/audit/*`
Model overrides and pricing	`MODEL_OVERRIDES_ENABLED=true`	edit model availability, metadata, and pricing in `Models`
Aliases and combos	dashboard `Models`, `Aliases`, `Combos`	expose stable model names and fallback combinations to clients
Provider pools	`pools:` YAML or dashboard `Pools`	route through pool names like `jina-pool/jina-embeddings-v3`
Workflows	`workflows.refresh_interval` and `/admin/api/v1/workflows`	enable or disable cache, audit, usage, guardrails, and fallback per provider/model/user path
Guardrails	`GUARDRAILS_ENABLED=true` and `guardrails.rules`	system prompt injection/override/decorator, LLM-based content altering, regex blocking, PII redaction, and length limits
Exact and semantic response cache	`cache.response.simple` or `cache.response.semantic`	reduce repeated provider calls; cache hits return `X-Cache` headers
Failover	`fallback:` YAML and `config/fallback.example.json` style rules	retry failed translated chat/responses calls on alternate models
Prometheus metrics	`METRICS_ENABLED=true`	scrape `/metrics`
Storage backends	`STORAGE_TYPE=sqlite`, `postgresql`, or `mongodb`	persist admin data, usage, audit, keys, policies, and caches

For the full Enterprise capabilities list with configuration details, see Enterprise Overview.

Run OSS

Fastest npm-based startup:

text

npx -y iaurora

Use the OSS example profile as a starting point:

text

$env:AURORA_CONFIG_PATH = "configs/editions/oss.example.yaml" $env:OPENAI_API_KEY = "<your-openai-api-key>" .\bin\aurora-oss.exe

Minimal Docker-style OSS configuration:

text

docker run --rm -p 8080:8080 \
-e AURORA_MASTER_KEY="sk-aurora-..." \
-e OPENAI_API_KEY="$OPENAI_API_KEY" \
-e LOGGING_ENABLED=true \
-e USAGE_ENABLED=true \
aurorahq/aurora

Then call the gateway:

text

curl http://your-aurora-host/v1/chat/completions \
-H "Authorization: Bearer sk-aurora-..." \
-H "Content-Type: application/json" \
-d '{
"model": "<model id returned by /v1/models>",
"messages": [{"role": "user", "content": "Hello from OSS"}]
}'

For Enterprise access, deployment, and capability checks, see Enterprise Overview.

Edition

Artifact

Best for

License

OSS

aurora-oss binary or public container

local teams, simple production gateways, provider aggregation, observability, and policy basics

not required

Enterprise

Provided by Aurora during onboarding

organizations needing identity, RBAC, tenant administration, budget enforcement, advanced routing, compliance, HA controls, and export connectors

license and access handled by Aurora

OSS features

These features are available without an Enterprise capability:

Feature	Configure	Use
OpenAI- and Anthropic-compatible gateway	provider env vars or `providers:` YAML	call `/v1/chat/completions`, `/v1/responses`, `/v1/embeddings`, `/v1/models`, `/v1/files`, `/v1/batches`, and `/v1/messages`
Provider auto-discovery	provider env vars such as `OPENAI_API_KEY`, `ANTHROPIC_API_KEY`, `GEMINI_API_KEY`, `OLLAMA_BASE_URL`, or `VLLM_BASE_URL`	list models at `/v1/models` and request them by model ID
Provider passthrough	`ENABLE_PASSTHROUGH_ROUTES=true` and `ENABLED_PASSTHROUGH_PROVIDERS`	call `/p/{provider}/...` for native provider APIs
Admin dashboard and API	`ADMIN_UI_ENABLED=true`, `ADMIN_ENDPOINTS_ENABLED=true`	open `/admin/dashboard` or call `/admin/api/v1/*` with the master key
Managed API keys	dashboard `Auth Keys` or `/admin/api/v1/auth-keys`	issue gateway keys without sharing upstream provider keys
Usage and cost analytics	`USAGE_ENABLED=true`	view `Usage` and `/admin/api/v1/usage/*`
Audit logs and live console	`LOGGING_ENABLED=true`	view `Audit Logs`, `Console`, and `/admin/api/v1/audit/*`
Model overrides and pricing	`MODEL_OVERRIDES_ENABLED=true`	edit model availability, metadata, and pricing in `Models`
Aliases and combos	dashboard `Models`, `Aliases`, `Combos`	expose stable model names and fallback combinations to clients
Provider pools	`pools:` YAML or dashboard `Pools`	route through pool names like `jina-pool/jina-embeddings-v3`
Workflows	`workflows.refresh_interval` and `/admin/api/v1/workflows`	enable or disable cache, audit, usage, guardrails, and fallback per provider/model/user path
Guardrails	`GUARDRAILS_ENABLED=true` and `guardrails.rules`	system prompt injection/override/decorator, LLM-based content altering, regex blocking, PII redaction, and length limits
Exact and semantic response cache	`cache.response.simple` or `cache.response.semantic`	reduce repeated provider calls; cache hits return `X-Cache` headers
Failover	`fallback:` YAML and `config/fallback.example.json` style rules	retry failed translated chat/responses calls on alternate models
Prometheus metrics	`METRICS_ENABLED=true`	scrape `/metrics`
Storage backends	`STORAGE_TYPE=sqlite`, `postgresql`, or `mongodb`	persist admin data, usage, audit, keys, policies, and caches

For the full Enterprise capabilities list with configuration details, see Enterprise Overview.

Run OSS

Fastest npm-based startup:

text

npx -y iaurora

Use the OSS example profile as a starting point:

text

$env:AURORA_CONFIG_PATH = "configs/editions/oss.example.yaml" $env:OPENAI_API_KEY = "<your-openai-api-key>" .\bin\aurora-oss.exe

Minimal Docker-style OSS configuration:

text

docker run --rm -p 8080:8080 \
-e AURORA_MASTER_KEY="sk-aurora-..." \
-e OPENAI_API_KEY="$OPENAI_API_KEY" \
-e LOGGING_ENABLED=true \
-e USAGE_ENABLED=true \
aurorahq/aurora

Then call the gateway:

text

curl http://your-aurora-host/v1/chat/completions \
-H "Authorization: Bearer sk-aurora-..." \
-H "Content-Type: application/json" \
-d '{
"model": "<model id returned by /v1/models>",
"messages": [{"role": "user", "content": "Hello from OSS"}]
}'

For Enterprise access, deployment, and capability checks, see Enterprise Overview.