Snapshot
Continue to About ↓- What it does
- xpander is a self-hostable control plane, runtime and gateway for enterprise AI agents.
More detail
- It lets a company run its agents, including Claude Code, Codex and OpenCode, in its own Kubernetes, VPC, on-premises or air-gapped environment, with every model call and tool call metered, budgeted, logged and tied to an owner.
- It is aimed at organizations that need AI agents inside their own security perimeter rather than in a vendor cloud.
- Best for
- Security and platform teams at regulated or sovereignty-constrained organizations (government, defense, critical infrastructure, banking, insurance, healthcare) that need agents, models and tools under one policy inside their own perimeter.
More detail
- Also fits teams already using Claude Code, Codex or OpenCode that want them off laptops with approvals and audit trails.
- Pricing
- Two commercial models, both quote-gated.
More detail
- Team is usage-based credits: 1 credit = $0.01 ($100 = 10,000 credits)
- 1 credit per agent wake for the whole turn, 1 credit per tool or API call, and model tokens at xpander-configured credit rates (published examples: Claude Sonnet 5 at 375 credits per 1M in
- Main limitation
- Buyers who want a self-serve free tier or quick card-swipe signup: onboarding is demo-first and the Enterprise option is an annual contract quoted from 50 agents.
More detail
- Poor fit where total cost must be known up front, since the Enterprise license fee is not published.
- Record
- Living and monitored
- Reviewed
- Evidence
- Public web research
- Machine access
- Structured profile ↗
About
- xpander is a self-hostable control plane, runtime and gateway for enterprise AI agents.
- It lets a company run its agents, including Claude Code, Codex and OpenCode, in its own Kubernetes, VPC, on-premises or air-gapped environment, with every model call and tool call metered, budgeted, logged and tied to an owner.
- It is aimed at organizations that need AI agents inside their own security perimeter rather than in a vendor cloud.
What it helps with
- Shadow AI: agents running on laptops or unattended with no shared policy or record of what they can reach.
- AI sprawl: a separate gateway per model provider, with no owner for models or spend.
- Ungoverned model spend: no per-person, per-group, per-agent or per-API-key budgets.
- Integration tax: buying an MCP gateway, an orchestrator and a control plane separately and stitching them together.
- Compliance boundary risk: vendor-hosted AI moving data outside the perimeter with no record of which agent touched what.
- Prototype stall: teams cannot say which agents are in production, what they may touch, or who approved them.
Capabilities
A practical view of what is available and where limitations apply.
| Offering | Available | Detail |
|---|---|---|
| AI Gateway | Yes | One gateway over commercial providers on your own keys, self-hosted model servers such as vLLM, and existing gateways such as Portkey or Kong. |
| Budgets and cost attribution | Yes | Budgets per person, group, agent and API key, daily, weekly or monthly, set to block calls or only alert; every call logged with model, tokens and cost under the requester. |
| Agent Runtime for coding agents | Yes | Runs Claude Code, Codex or OpenCode in an isolated sandbox that receives a gateway token, never a provider key; scheduled agents persist after the laptop closes. |
| MCP Gateway | Yes | Agents connect only to approved tools, with each person signing in as themselves. |
| Agentic Workflows with approvals | Yes | Multi-step processes that pause for a named human approver where you decide. |
| Self-hosted and air-gapped deployment | Yes | Managed cloud, your VPC, your Kubernetes, on-premises or fully disconnected; no data egress in self-hosted mode per the vendor. |
| Secrets isolation | Yes | Credentials held in a vault and injected at tool-call time after the model context ends, so the model never sees them. |
| Audit export | Yes | Runs, tool calls, approvals and costs recorded as they happen; audit log downloadable as CSV or JSON. |
| No-code agent builder | Yes | Omni, the vendor's no-code builder, plus an SDK and REST API for code-first agents. |
| Independent third-party reviews | Unknown | No third-party review or analyst corpus was located in the evidence gathered; certifications are vendor-stated pending the SOC 2 report via the Trust Center. |
Bot trust
How confidently AI systems can verify this profile from public evidence.
—of 100
ChatGPT—/100
Claude—/100
This profile has not completed a Bot Score assessment yet.
Evidence
Where this record came from, when it was reviewed, and how confidently it can be used.
- Verification
- Scribble researched these facts from the linked public sources.
- The owner has not manually confirmed every field.
- Last reviewed
- Primary source
- xpander.ai
Technical profile details
Record ID
3d3276d7-b4c5-4aa1-8822-a0f9fe4bca81Canonical URLhttps://profiles.scribble.network/listing/xpander-3d3276d7