Enterprise-ready, Cloud or on-prem

One gateway for every
model and tool call.

Appwire sits between your AI clients, your models, and your MCP servers. It controls model access, usage, and cost, and makes every tool call authenticated, authorized, approved, and audited, behind one URL your whole organization can use.

Works with your model providers and any MCP client

AI clients
Models
AI clients connect through the Appwire gateway to MCP servers CClaude Desktop >_Claude Code CuCursor VSVS Code SDKCustom agents OAOpenAI ANAnthropic GEGoogle Gemini MCP SERVERS GHGitHub JIJira PGPostgres appwıre AI + MCP GATEWAY Models Usage Cost SSO Policy Approve Secrets Audit OTel 3 providers · 12 servers · 1 URL
Authenticated jane@acme.com via Okta · platform-eng
Approval requested jira.create_issue · sent to #eng-approvals
Blocked postgres.drop_table · policy deny *.drop_*
Budget alert data-science at 85% of today's model budget

Drop-in for the models, clients, and servers you already use

The problem

AI adoption is outrunning AI governance.

Teams now wire up models and MCP tools in minutes. Nobody can see who uses which model, what it costs, who can call which tool, or what happened next.

No control over model spend

Every team calls models with its own API keys. Nobody can say who uses which model, and the spend shows up only on the monthly bill.

Servers everywhere, inventory nowhere

Every team runs its own MCP servers with credentials pasted into local config files. Nobody can list what is deployed, let alone what it can do.

No identity at the tool layer

Tools execute as whatever token happened to be in the client, not as the person or agent behind the request. Least privilege is impossible.

Zero visibility after the fact

When an agent updates the wrong record, the only log lives on a laptop. Try explaining that to your auditor, or your customer.

AI gateway

Control models, usage, and cost.

Point your apps and agents at Appwire instead of at each provider. Decide which models every team can use, cap how much they use, and see exactly where the spend goes.

Model access

Allow or deny providers and models per team, app, or key, in the same policy that governs your tools. Provider keys stay in your vault.

Usage limits

Rate limits and quotas per user, team, or key, so one runaway agent cannot use up everyone else's capacity.

Cost control

Budgets and alerts per team, with every dollar of model spend attributed to the person, agent, and team behind it.

Model spend · todaylive
$912.4061% of the $1,500 daily budget, across 3 teams
  • platform-eng
    $412 of $700
  • data-science
    $386 of $450
  • support-copilot
    $114 of $350
Anthropic 52%OpenAI 35%Google 13%

data-science passed 85% of its daily budget. Alert sent to #ml-leads.

MCP gateway

Govern, secure, and observe every tool call.

Appwire is an MCP server to your clients and an MCP client to your servers. That position lets it enforce identity, policy, and audit on every request without changing either side.

Policy-based access control

Roles sync from Okta, Microsoft Entra ID, Google Workspace, or any OIDC provider. Allow, deny, or require approval per server, per tool, or per argument pattern. One policy, enforced everywhere.

policy: platform-eng-default
subjects: { groups: [platform-eng] }
rules:
  - allow: github.*   where: { repo: "acme/*" }
  - require_approval: jira.create_issue
  - deny: "*.delete_*"

Observability and immutable audit

Every call is logged with the principal, tool, argument hash, latency, cost, and the parent LLM trace. Export logs, traces, and metrics over OpenTelemetry to Datadog, Grafana, Splunk, or BigQuery.

auth3 ms
policy2 ms
secrets5 ms
upstream212 ms
record1 ms

Virtual keys and tool groups

Scope access per user, team, environment, or customer. Attach rate limits and budgets to each key. Rotate without touching a single client config.

Upstream OAuth and secrets

Appwire holds the credentials, not your laptops. OAuth 2.0 at the server or per-user level with automatic refresh, backed by Vault, AWS, GCP, or Azure secret stores.

Human-in-the-loop approvals

Mark tools read-only, auto-approved, or approval-required. Approvers get the full request in Slack, Teams, or email. Nothing runs until someone says so.

Virtual MCP servers

Bundle tools from many upstream servers into one curated server. Clients see one URL and only the tools they are allowed to use.

Guardrails in both directions

Scan arguments and results for secrets, PII, and prompt injection. Redact or block before anything reaches the model or the server.

Token-efficient tool loading

Load schemas on demand instead of shipping hundreds of tool definitions in every prompt. Smaller context, faster runs, lower bills.

How it works

One tool call, traced end to end.

Select any stage to see exactly what Appwire does before, during, and after a tool executes.

trace 7f3a…c21e 01 / 06

Control plane

See everything from one console.

Inventory every model, server, and tool, approve or deny requests, trace every call back to the person and agent behind it, and attribute cost to the team that owns it.

console.appwire.io / acme / overview live
Model and tool calls · 24h52,908+12% vs yesterday
p95 latency118 msgateway overhead 4 ms
Pending approvals4oldest 6 min
Spend today$912.40
61% of $1,500 budget
TimeCallPrincipalUpstreamLatencyStatusCost
14:02:31github.create_pull_requestjane@acme.comgithub231 msallowed$0.0004
14:02:30anthropic/claude-sonnetagent:support-copilotanthropic1,840 msallowed$0.0213
14:02:29slack.post_messageagent:release-botslack96 msallowed$0.0002
14:02:26jira.create_issuesam@acme.comjira—approval—
14:02:24openai/gpt-5priya@acme.comopenai2,215 msallowed$0.0318
14:02:22postgres.queryagent:analyticspostgres188 msallowed$0.0011
14:02:19postgres.drop_tablecontractor@vendor.iopostgres2 msblocked$0
14:02:15salesforce.update_opportunitypriya@acme.comsalesforce342 msallowed$0.0008

Quick start

Up and running in minutes, not quarters.

Run the gateway, add your model providers and MCP servers, point a client. Add SSO, policy, and approvals when you are ready. No changes to your servers, no SDK to adopt.

  • STDIO, streamable HTTP, and SSE transports
  • Single binary, Docker image, or Helm chart
  • Config as code, with a UI when you want one
  • Import existing client configs in one command
shell
# Run locally
$ npx appwire@latest

# ...or with Docker
$ docker run -p 8080:8080 ghcr.io/appwire/gateway:latest

# ...or on Kubernetes
$ helm install appwire oci://ghcr.io/appwire/charts/appwire

✓ gateway listening on http://localhost:8080/mcp
✓ console at           http://localhost:8080

Deployment and compliance

Your infrastructure. Your keys. Your rules.

Appwire runs where your data lives, with no outbound dependencies. Credentials never leave your environment, and every action produces audit-ready evidence.

Self-host in your VPC

Helm chart and Terraform modules for AWS, GCP, and Azure. Horizontal scaling with cluster mode, using Postgres or your managed database as state.

On-prem and air-gapped

A single binary with offline licensing and zero call-home. Bring your own identity provider, secret store, and observability stack.

Appwire Cloud

Fully managed, single-tenant deployments with regional data residency. Same binary, same policies, none of the operations.

Built for regulated environments

Immutable audit logs, SCIM provisioning, encryption in transit and at rest, and exportable evidence for your SOC 2 Type II, HIPAA, GDPR, and ISO 27001 programs.

  • SOC 2 Type II
  • HIPAA
  • GDPR
  • ISO 27001

Editions

Built for growing teams. Ready for the enterprise.

Every plan includes the full AI and MCP gateway, every transport, and unlimited servers. Enterprise adds the governance layer large organizations need.

CapabilitySmall to Medium businessEnterprise
AI and MCP gateway core, all transports, unlimited servers✓✓
Virtual MCP servers and tool groups✓✓
OpenTelemetry export✓✓
Policy for models and tools, with approvalsBasic✓ Advanced, argument-level
SSO (SAML / OIDC) and SCIM—✓
Secret store integrations (Vault, AWS, GCP, Azure)—✓
Guardrails: PII, secrets, prompt injection—✓
Cluster mode and high availability—✓
Budgets, alerts, and cost attribution—✓
Audit log retention7 daysCustom, with legal hold
DeploymentAppwire Cloud or self-hostedVPC, on-prem, air-gapped, or Appwire Cloud
SupportEmail, next business day24×7 with SLA and dedicated engineer

Pricing

Start free. Scale with governance.

No per-call fees. Pay for the people and teams you govern, not for how much your agents work.

Small to Medium business

$0up to 5 users

For growing businesses connecting their first AI tools to the systems they already run on.

  • AI and MCP gateway, unlimited servers
  • Virtual servers and tool groups
  • Basic policies and approvals
  • OpenTelemetry export
  • 7-day audit retention
  • Email support
Get started free

Enterprise

Customannual agreement

For regulated organizations that need Appwire on their own terms.

  • Everything in Team
  • VPC, on-prem, and air-gapped deployment
  • Secret store integrations and guardrails
  • Cluster mode and high availability
  • Custom retention with legal hold
  • 24×7 support with SLA
Talk to sales

FAQ

Questions, answered.

What is an AI and MCP gateway?

An AI gateway sits between your apps and agents and the model providers they call, so you control which models are used, by whom, and at what cost. An MCP gateway does the same for tools: it sits between MCP clients (Claude Desktop, Cursor, your own agents) and MCP servers (GitHub, Slack, your databases), and handles authentication, authorization, credentials, and logging for every tool call. Appwire is both, behind one endpoint.

Do I need to change my clients, models, or MCP servers?

No. Appwire speaks standard MCP on both sides, so you register your existing servers over STDIO, streamable HTTP, or SSE and point your clients at one Appwire URL. Model providers are added the same way, with their keys held in your vault. Servers and models keep working exactly as they do today.

Does Appwire execute tools automatically?

Only when your policy says so. Each tool can be marked auto-approved, approval-required, or denied, and rules can depend on the caller, the arguments, or both. By default, write operations require explicit approval and read operations are allowed.

Where do credentials live?

In your environment. Upstream OAuth tokens, model provider keys, and other API keys are stored in your secret manager (HashiCorp Vault, AWS Secrets Manager, GCP Secret Manager, Azure Key Vault) or Appwire's encrypted store, and injected at request time. Clients only ever hold a revocable virtual key.

Can Appwire run without internet access?

Yes. The gateway has no outbound dependencies and supports offline licensing. Air-gapped deployments are a first-class use case for the Enterprise edition.

Do I still need a separate LLM gateway?

No. Appwire controls model access, usage, and cost the way an LLM gateway does, and governs tool calls the way an MCP gateway does, with one identity, one policy, and one audit trail across both. Every tool call carries the trace of the model call that asked for it.

What is included in each plan?

The free Small to Medium business plan includes the full AI and MCP gateway core, all transports, virtual servers, basic policies, and OpenTelemetry export for up to five users. Team adds SSO, advanced policy, budgets and cost attribution, and 90-day audit retention. Enterprise adds secret store integrations, guardrails, cluster mode, VPC, on-prem, and air-gapped deployment, and custom retention.

How much latency does the gateway add?

Authentication, policy evaluation, and audit logging run in-process and typically add single-digit milliseconds to a model or tool call. Upstream execution time is unchanged and reported separately so you can see exactly where time is spent.

Get started

Put every model and tool call behind one gateway.

Tell us about your environment and we will set up a 14-day Enterprise trial on your infrastructure, or walk you through the gateway in a 30-minute demo.

  • Live demo tailored to your models and MCP servers
  • Architecture review for VPC or air-gapped rollout
  • No credit card, no commitment

Prefer email? hello@appwire.io

We reply within one business day. By submitting you agree to our privacy policy.