A AMCP ·
Powered by GMI
connecting… /api → :8080

Mission control for your self-hosted agent

A self-hosted coding agent running on GMI Cloud AgentBox, wired to GMI MaaS for inference. This page watches every session live — status, tool calls, and arguments as they happen — and can drive them too.

sessions working right now
Sessions
persisted on the server
Messages exchanged
across all sessions
Tools available
– agent types
Server uptime
protocol –

GMI MaaS wiring

model: –

Every setting has a working default, so this runs with no environment variables at all. Each one below is an override you can set if you want to change something.

1

Works out of the box

Defaults are baked into the image: a shared trial key and DeepSeek V4 Flash, against GMI MaaS.

model    = "deepseek-ai/
            DeepSeek-V4-Flash-0731"
base_url = "https://api.
            gmi-serving.com/v1"
2

Models are slugs

The chat API identifies models by slug. AgentBox's GMI_MODELS UUID is a different namespace and is ignored — it returns 404 here.

GMI_MODEL="deepseek-ai/…"  ✅
GMI_MODELS="7f3d…-uuid"    ❌ 404
3

Override anything

All optional. Or change the key and model live from the panel below — no restart.

GMI_API_KEY=…    own key
GMI_MODEL=…      own model
AMCP_USE_AGENTBOX_MODEL_ID=1
                 use the UUID
Model source
AMCP_CHAT_MODELGMI_MODEL → default slug (GMI_MODELS only if opted in)
Key source
GMI_API_KEYGMI_MAAS_API_KEYOPENAI_API_KEY → bundled trial key
Wire format
OpenAI-compatible chat completions, tool calling supported
Topology
gateway on :8080 (public) → amcp serve on 127.0.0.1:8081

API key

checking…
Get your API key

This deployment ships with a shared trial key so it works immediately. After the free prompts are used, add your own key to keep going — it is written to the running config and takes effect on your next prompt.

Provider status unknown — run the connection test to check the key and model end to end.

The test creates a throwaway session, sends a one-word prompt, and reads the result — the only way to confirm the key and model slug actually resolve. Failures surface as provider errors rather than crashes, so it is safe to re-run. Note that it spends one free prompt while the trial key is in use.

Sessions

No sessions yet.
Create one to start.

Live activity

no session selected
idle
Pick a session on the left to attach to its event stream.
Attach to a session first

Tool calls in this session

No tool calls observed yet on this session.

Counts are observed live from this session's event stream, so they start at zero each time you attach. Bars share one hue because this encodes magnitude, not identity — the tool names on the axis carry identity.

Runtime capabilities

Built-in tools

loading…

Agent types

loading…

Start creating today

Spin up your own agents on GMI Cloud — bring a key, pick a model, and go from a blank workspace to a running coding agent in minutes.

Open GMI Cloud console