1m0.0
5m1300.0
1h2174.8
24h93.2
Rolling windows; the 1m value includes live in-flight activity when available.
Updated just nowOrchestris Server
One control plane for every AI request.
Connect providers once, publish the right models to each team, and give both people and agents a governed route to AI. Orchestris Server keeps credentials, policy, usage, cost, health, and audit evidence in the environment you choose.
Community Edition is free for one organization with up to 8 users. No license key required.
Interactive product replica
Admin UI, ready to explore.
Govern providers, models, people, usage, and server settings.
▣Aster Ridge Operations
⌕ Search /mara.wells@asterridge…Dashboard
Organization Metrics
Org: Aster Ridge OperationsToday: 2026-07-09 (America/Chicago)
Quick actionsCreate group
Today2026-07-09$222.15
MTDSince 2026-07-01$1,805.30
Peak Active (within 15m)20
WS users(now)—WS connections(now)0
Provider PerformanceProvider cost rows reflect provider-attributed daily cache data and may not sum exactly to the headline estimated cost when usage cannot be attributed to a provider.
Active Providers4Requests (1h)3,060Errors (1h)0Error Rate (1h)0.00%
Provider Health No provider failures in 30d
ProviderCost TodayCost MTDAvg PP (Day)Slowest PP (Day)Avg Gen (Day)Slowest Gen (Day)TPS (15s)24h TPS
› OpenAI Business GatewayUpdated 1m ago$62.23$505.51115.071.892.055.2--0.3
› Anthropic Claude GatewayUpdated 1m ago$46.68$379.1495.059.376.045.6--0.3
› Google Gemini GatewayUpdated 1m ago$37.79$306.92110.068.688.052.8--0.3
› Private Model ClusterUpdated 1m ago$75.57$613.84170.0106.1136.081.6--0.5
7-Day Tokens by Provider
Org-scopedToday (live) Private Model Cluster 3.3B Google Gemini Gateway 453.4M OpenAI Business Gateway 419.6M Anthropic Claude Gateway 89.5M
2026-07-032026-07-042026-07-052026-07-062026-07-072026-07-082026-07-09
What to try
1Click Orchestris Admin for Dashboard
2Open Directory or Integrations
3Review Reports and Settings
Static website demo · simulated Aster Ridge data · no product systems are connectedControl without lock-in
Operate the access layer between your teams and your models.
Orchestris keeps provider-specific details behind a consistent control plane, while preserving the identity, ownership, and usage context operators need.
01
Provider and model control
Connect commercial and private providers, discover their catalogs, publish stable model aliases, and decide which groups can use each model.
02
Identity and access policy
Manage organizations, members, roles, groups, API keys, model grants, quotas, and request limits from one administration surface.
03
Usage and cost evidence
Review request volume, tokens, latency, provider health, estimated cost, quota use, and auditable administrative changes.
04
Client sync and storage
Back Orchestris Chat with authenticated WebSocket streaming, conversation sync, projects, attachments, and multi-device continuity.
05
Agent API gateway
Give compatible tools one OpenAI-style base URL while applying the same model access, keys, quotas, aliases, and reporting used by people.
06
Private deployment choices
Start on a single machine, deploy a container or native package, and move to PostgreSQL and shared infrastructure as operating needs grow.
Client integration
Give the client a home, not just an endpoint.
Orchestris Chat signs in to the server, receives an organization-scoped model catalog, and streams responses over an authenticated connection. Conversation state then follows the user across supported devices.
- No personal provider keys required for server-managed models.
- Multi-device conversation and message synchronization.
- Server-backed projects, attachments, titles, and deletion state.
- Model availability and quotas follow organization policy.
OpenAI-compatible agent gateway
Bring scripts and agents through the same policy layer.
Compatible SDKs and tools can use /openai/v1 as their base URL. Calls resolve through configured providers and model aliases, then arrive in the same usage and cost reports as client traffic.
- Chat Completions, Responses, and Models compatibility routes.
- Organization keys for services and user keys for individual tools.
- Per-key model allowlists, budgets, lifecycle controls, and attribution.
- Stable aliases let operators change backing models without redeploying clients.
Gateway Operations
Agent access, policy, and usage in one workflow
Configure keys, route models, monitor usage, and manage providers from the same Orchestris control plane.
User API Keys
Members can create user-owned keys with model allowlists, reveal and rotate controls, and per-key budget, used-today, and remaining quota tracking. Org admins use the organization API Keys screen for shared service keys.
Model Aliases
Agents call stable aliases while admins retarget backing models without redeploying agent code.
Usage Reports
Agent traffic shows up in usage reporting with tokens, latency, estimated cost, models, and providers.
Provider Catalog
The gateway routes through configured providers instead of hard-coding one vendor in each agent.
Included control surface
The pieces required to run AI as a shared service.
Start with the full operating loop: connect supply, grant access, serve requests, and explain what happened afterward.
Provider catalog and model discovery
Model aliases and group-based grants
Organization and user administration
User-owned and organization-owned API keys
Quota, rate-limit, and concurrency controls
Usage, cost, health, and audit reporting
Chat streaming, sync, projects, and attachments
OpenAI-compatible chat, responses, and model routes
Free self-hosted start
Community Edition
Run Orchestris Server on one machine with SQLite for one organization and up to eight users. Install a native package or Docker image and keep the data on your infrastructure. No license key is required.
WindowsmacOSLinuxDockerSQLite
View server downloadsGrowing organizations
Commercial deployment
Expand into PostgreSQL-backed, multi-node, or dedicated managed environments with larger user limits, rollout planning, and support for your operating model.
PostgreSQLMulti-nodeDedicatedSelf-hostedSupport
Discuss your deployment