Choose theme
AI Resources
Magpie
Magpie manages the models and providers used by supported coding agents from one application.
It connects agents to a local gateway and offers a shared model picker, routing groups, saved profiles, and usage views. The project provides desktop and terminal interfaces for macOS, Windows, and Linux. Use this as a first read, not a recommendation. Open the original project before trusting details like terms, limits, privacy, cost, setup, or safety.
What it is
A model gateway for coding tools
Magpie sits between supported agents and the providers you configure. It updates each connected agent's endpoint and model settings, then routes requests through its gateway.
Why it stands out
One model choice across several agents
The picker draws on configured providers, including API services and local servers such as Ollama or LM Studio. A saved profile can hold several agents' model choices together.
Availability
Downloads and public implementation
The project declares an MIT license and provides downloads, Go source, Docker instructions, and a command-line interface. Its documentation also covers a terminal UI, web UI, and Termux.
Why it matters
What makes it useful
Someone using several coding agents can otherwise end up maintaining a different model name, endpoint, and key in each tool. Magpie brings those choices into one picker. For example, it documents choosing a DeepSeek model for Claude Code and a Kimi model for Codex. Profiles let you save a set of choices and return to it later, rather than reconstructing every agent's setup.
What to know
Where it fits
Start with an installed agent, add a provider, and choose the model that agent should use. Magpie shows the supported agents it detects and writes the relevant configuration in each agent's own format. Its gateway accepts OpenAI Chat and Responses, Anthropic Messages, and Gemini protocols. The selected upstream provider still receives the requests; a local gateway does not turn a remote model into a local one.
Notable points
What stands out
A routing group can appear as one model choice while containing several models or accounts. The documented order mode tries the first available member before moving on; rotate spreads turns across members, while other modes consider usage or quota windows. This means a group name can hide a change in the model answering a request. The Routing view exposes the decisions, and the usage views show tokens, quota, and estimated cost.
Before using
What to review
Magpie changes connected agents' configuration. Its documentation describes recording previous settings for disconnection and detecting later configuration drift; some agents need a restart or new session to read changes.
Subscription discovery reads existing agent sign-ins and supported credential stores. The README describes a Subscription accounts privacy switch and an accounts discovery off command; API-key providers remain available with discovery disabled.
Released builds report a daily usage event with an installation ID, version, system, and configured agent, provider, and model identifiers. The documentation provides controls to disable it, and separately describes website partner-link counting and icon fetching.
Provider access, quotas, costs, and agent compatibility still depend on the selected services and models. Magpie's cost views include estimates based on model pricing; they are not a promise of a particular bill or uninterrupted access.
Reader fit
Who may find it relevant
Relevant to developers managing more than one coding agent, model provider, or account. The desktop picker offers a way to manage those settings without repeatedly editing configuration files. If one agent on its usual provider already meets your needs, the extra gateway and configuration management may add little.
Editorial note
Why LifeHubber lists it
We list Magpie because it gives people using several coding tools a common place to manage the model connections behind them. That separates two choices that are easy to blur: the agent you work in and the model that answers it. A familiar agent can remain part of your workflow while you consider a different configured provider; the new model's tool support and responses still need their own assessment. This makes the project relevant to someone who wants to change a model connection without choosing a whole new coding interface at the same time.
Source links
Source materials
Reader note
Before relying on this entry
LifeHubber lists entries to help readers inspect AI projects, not to endorse them or prove they are safe, suitable, accurate, maintained, or right for a specific use. We do not verify every entry in depth. Before relying on anything listed, review the original materials, terms, privacy practices, limits, and risks that matter for your situation.
What to explore next
Plan the work around a model change.
Changing the connection is one part of switching models. Continue with a guide to finding workflow dependencies and checking a replacement against the work you need it to do.
More in AI Agents
Keep browsing this category
Explore more AI agent projects.
Paperclip
paperclipai/paperclip
A self-hosted server and dashboard for coordinating agent teams through companies, goals, roles, issues, heartbeats, budgets, approvals, and persistent activity records.
Open Instinct
mariagorskikh/open-instinct
A beta personal-agent project with local CLI chat, messaging and app integrations, readable memory, and permission tiers for requests from other people and agents. Its self-hosted messaging tunnel has a reported delivery problem with a fix proposed.
Microsoft Agent Governance Toolkit
microsoft/agent-governance-toolkit
A Microsoft toolkit for adding policy enforcement, identity, sandboxing, audit records, reliability tooling, and the Agent Control Specification runtime layer around AI agents, with Python, TypeScript, .NET, Rust, and Go surfaces.
For project maintainers
Listed here? You can use the badge.
If you maintain a project with a current LifeHubber listing, you may add the optional “Listed on LifeHubber AI Resources” badge to its README, docs, or website. No introduction or permission request is needed.
