Build complete applications in the terminal with an AI agent
Combine models such as Qwen to iteratively write and refine code step by step via the agent
Best for：Independent developers
Coding / Developer
omp
omp is an open-source terminal coding agent that supports over 60 model providers and allows switching models within a session.
Product type: CLI Coding AgentPricing: UnknownAPIMaster integration: Supported
Website created
2026-05-11
OpenRouter token usage rank
#7
Source: OpenRouter
Monthly unique visitors
187,532
Source: SimilarWeb
Coding Plan
Stop juggling model subscriptions
Coding Pro model discounts
- glm-5.250% off
- glm-5.350% off
- deepseek-v4-flash40% off
- deepseek-v4-pro40% off
- qwen3.8-max40% off
- kimi-k2.7-code30% off
- kimi-k330% off
- minimax-m330% off
Overview
omp is positioned as a command-line agent tool for developers to use in local projects, running via TUI or GUI interfaces. After launching in the project directory, users can have the agent perform CLI operations and invoke different models to handle tasks. The core feature is support for mid-session model switching and local configuration, avoiding single-model limitations. It suits developers needing flexible model routing and handling real codebases, emphasizing local persistence and multi-model switching compared to cloud-shared agents or fixed-model tools.
One-line summary
omp is an open-source terminal coding agent that supports over 60 model providers and allows switching models within a session.
What people use it for
After users enter omp in the terminal to launch, the agent uses CLI to complete AWS/GCP operations or code modifications; when encountering model limits, enter /switch to switch to another model and continue the task.
Best for
Developers and technical teams
How it works
Run the CLI in the project directory, configure model providers and endpoints, then approve or inspect the agent's operations in the terminal.
Quick facts
Product type
CLI Coding Agent
Pricing
Unknown
Please check the official repository for the project's current pricing and model invocation rules.
APIMaster integration
Supported
~/.omp/agent/models.yml or model Provider configuration
Data confidence
Low
Last verified: 2026-08-29
Supported models / APIs
Provider configuration depends on versionCustom endpoint support: to be verified
Core features and differentiators
Native LSP Integration
Directly provide semantic renaming, reference lookup, and diagnostics within the terminal
Full DAP Debugger
In-process debugger supporting breakpoints, variable inspection, and step execution
Hashline Content Hash Editing
Content-hash-based code modification mechanism
Persistent IPython Kernel
Retains imports, variables, and open file state across sessions and sub-agents
Browser Automation
Built-in browser operation capabilities
Multi-model Provider Support
Use the same toolset to invoke different models via TUI, -p, RPC, ACP and other interfaces
What is omp best for?
Debug Python programs in an IDE-integrated environment
Use the native DAP debugger to set breakpoints and inspect variables
Best for：Python developers
Incrementally build state like a Jupyter notebook
Accumulate variables and file state step by step via the persistent kernel for iterative development
Best for：Data scientists and script developers
Execute automation tasks involving browser interaction
Have the agent directly operate the browser to complete web-related tasks
Best for：Full-stack developers
Third-party key setup for omp (APIMaster integration)
Setup steps
- 1Open omp's model Provider configuration as per official instructions, usually by editing ~/.omp/agent/models.yml.
- 2Add a custom OpenAI-compatible provider, fill in APIMaster's Base URL and your API Key.
- 3Change the model name to the APIMaster model ID you want to call and save the configuration.
- 4After restarting or refreshing the omp session, switch to that provider and use a minimal request once to confirm the current model is using the third-party key.
Config values
Base URL
https://apimaster.ai/v1API key environment variable
$APIMASTER_API_KEYModel
Your APIMaster model IDCommunity discussion
Discussion summary
Users typically understand omp as an IDE-integrated AI coding agent tool, emphasizing persistent state management and model routing features for local development and agent workflows. Discussions focus on comparisons with other agent tools, especially memory persistence, orchestration reliability, and real-world delivery capability. Users also pay attention to model routing applications across different tasks and integration methods with specific providers.
Most discussed topics
- 1
How the persistent IPython kernel supports incremental state building by agents
Users discuss how this design retains variables, imports, and files across sessions, enabling step-by-step iteration like a Jupyter notebook rather than re-deriving everything each time.
- 2
Real-world effectiveness of model routing by task type or difficulty
Users share how routing automatically selects models based on tasks and its impact on coding efficiency and cost.
- 3
Comparison with other agent tools such as Droid in terms of memory and orchestration
Users compare omp's durable harness and shared agents, discussing which is better suited for real project delivery and feedback loops.
- 4
Integration and usage experience with providers such as Google
Users inquire about configuring providers, anti-gravity subscriptions, and other details, as well as omp's stability in production environments.
- 5
Overall reliability and suitable scenarios of omp as a local coding agent
Users discuss its advantages in personal application development and avoiding cloud restrictions, as well as pairing methods with Claude or other models.
FAQ
What is omp?
It is a terminal-oriented Coding Agent project; specific capabilities should be verified against the current state of the repository.
Does it support APIMaster?
Currently unknown. The page will not claim support until confirmed in official documentation or actual testing.
Can it modify local files?
This depends on the enabled tools and permissions; each operation should be checked before execution.
Similar agents
Sources：Official websiteGitHub
Last verified: 2026-08-29 · Report a correction