Idun SDK

Thin, stdlib-only client + CLI for the NatureLM-Idun-5-MoE agent on Azure AI Foundry.

No httpx, no azure-identity, no Flask — it runs headless on Termux/Android with nothing but the Python standard library. Idun is a tool agent (it reasons and calls tools like web_search); this SDK surfaces the full agent trajectory (reasoning steps + tool calls) instead of a black-box answer.

---

Install

pip install idun-sdk

Installs the idun CLI, the idun Python package, and the stdlib MCP server idun_mcp.py.

Authenticate

idun login          # device-code flow, token saved to ~/foundry_token.txt

No admin role required — idun login returns a plain Entra bearer token for the Foundry endpoint. Any tenant user with agent access (RBAC) can run it; admin rights are only needed to configure the agent, not to use it. The token auto-rotates before expiry.

Quickstart

# final answer only
idun chat "Summarize Contoso's sustainability communications in one sentence."

# full agent trajectory (reasoning + tool steps)
idun trace "Use web_search to find the current CEO of Contoso."

from idun import IdunClient

res = IdunClient().complete("Your prompt here")
print(res.text)                 # final answer
for s in res.steps:             # agent trajectory
    print(s.kind, s.tool, s.query, s.status)

Features

Command	Purpose
idun chat	final answer only
idun trace	full trajectory (steps + text)
idun export	trajectory as JSON / Markdown
idun packs / idun run	curated prompt packs (e.g. Contoso)
idun diff	side-by-side trace diff of two prompts
idun token	inspect / refresh the Foundry token
idun logo	print the Foundry logo

MCP server

idun_mcp.py is a zero-dependency stdio MCP server (no FastMCP / httpx):

python3 idun_mcp.py

Exposes idun_chat(prompt) and idun_trace(prompt). Add to any MCP client:

{ "mcpServers": { "idun": { "command": "python3", "args": ["/abs/path/idun-sdk/idun_mcp.py"] } } }

Docs mirror (GitMCP): https://gitmcp.io/qapdex-maker/idun-sdk/sse

Combine with Sentry (MCP)

Idun pairs well with Sentry's MCP server so an agent can both call Idun and inspect Sentry errors/traces. Sentry MCP is a separate, optional provider — the Idun SDK stays stdlib-only.

Remote (zero-setup, OAuth — note the /mcp suffix):

{ "mcpServers": { "sentry": { "url": "https://mcp.sentry.dev/mcp" } } }

Stdio (OAuth via mcp-remote, no static token needed):

{ "mcpServers": { "sentry": { "command": "npx", "args": ["-y", "mcp-remote@latest", "https://mcp.sentry.dev/mcp"] } } }

Authenticate once (opens a browser for the OAuth flow):

/mcp auth sentry

Recommended combo for a foreign agent: idun (calls the agent) + idun-docs (reads the SDK docs) + sentry (queries error tracking). The agent can invoke Idun and, when a call fails, pull the correlated Sentry issue on its own.

Request shape (verified)

POST {base}/api/projects/{project}/agents/{agent}/endpoint/protocols/openai/responses?api-version=2025-05-15-preview
Authorization: Bearer ***
{"model": "model-router", "input": "<prompt>", "max_output_tokens": 4096}

• model MUST be "model-router" (agent id is in the URL).
• Do not send a tools key — the agent owns its tools (else 400 invalid_payload).

Links

• PyPI: https://pypi.org/project/idun-sdk/
• Repo: https://github.com/qapdex-maker/idun-sdk
• Foundry: https://ai.azure.com