Dexto CLI
A CLI runs a coding agent with files, shell commands and memory, while YAML configuration defines the model and connected tools.
- 659 ★ on the repository
- Last commit
Broad access
Why this level
- The agent can edit code, run shell commands and connect external tools with process permissions.
- A selected cloud provider receives project context and consumes account quota or funds.
Install
Manual install
curl -fsSL https://dexto.ai/install | bashInstalls Dexto on macOS/Linux/WSL. Read the installer before running it, then configure the selected model provider.
Access and checks
How we reviewBroad accessThe tool can change data, run commands or handle secrets.
- The agent can edit code, run shell commands and connect external tools with process permissions.
- A selected cloud provider receives project context and consumes account quota or funds.
This is third-party code. Review the repository files before installing.
What it does
Dexto CLI provides a ready terminal interface for an agent working on a project. The agent reads and edits code, runs checks and retains conversation context. The default launch loads the ready-made `coding-agent`. The same runtime supports specialized agents defined in YAML. Separate workers and planning help with more complex tasks. Web UI and server operation are other platform surfaces; the CLI is sufficient for an initial local task.
Who it is for. For developers who want a configurable terminal coding agent with persistent sessions.
Good fit when
- Work on code with file inspection and test execution.
- Use project-specific model and MCP tool configuration.
- Resume context from an earlier session.
- Reuse one agent runtime for several specializations.
Not a fit when
- A model must not execute commands on the work computer.
- Project source code must not be sent to the selected cloud provider.
- You expect a ready service without provider and permission setup.
Example request
Prepare Dexto for a test project and first inspect the code structure and verification plan without auto-approval.Limitations
Connect a model with your own API key or supported cloud credentials; installing the CLI does not supply access. The selected provider can receive source code and tool responses, so check permission to share project material. `--auto-approve` disables action confirmations; leave it off during the first run on your files. Dexto sources use Elastic License 2.0, while dependencies and external services have separate usage terms.
How to disable. Exit the CLI and stop any separately running servers. Inspect saved sessions and project files before removing local Dexto data.
FAQ
What starts by default?
The included coding agent, also selectable with `dexto --agent coding-agent`.
Can I submit a one-off task?
Yes. `-p` accepts a task prompt for a single run.
How do I select terminal mode explicitly?
Use `dexto --mode cli`. `dexto --help` lists the other options.
How do I find model-selection commands?
Type `/` in the interactive CLI. The command list includes model switching, session management and tool configuration.
Which CLI version was inspected?
The CLI manifest specifies 1.14.4, and the project remains in beta. The unpinned installer may deliver a newer build over time.
Discussion
No questions or reviews yet.
Related
CLI toolsDreamCoder’s own CLI runs coding tasks in a local workspace using the model provider selected by the user.
A CLI agent that works with project files and shell commands, lets you choose a model provider and connects external MCP tools.
AgenticX CLI comes with the base Python package; developers add memory, document processing and OCR as separate dependency groups.
A terminal agent explores projects, prepares plans and makes edits or runs commands under a chosen approval policy.
OpenAI's coding agent that runs locally in your terminal, with IDE and desktop versions
A CLI connects installed coding agents into a team, dispatches work between sessions and collects results across machines.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Dexto CLI (this card) | no | 659 | |
| OpenCode | no | 224,828 | a terminal coding agent with separate build and plan modes |
| EDDI CLI | no | 390 | EDDI’s CLI updates a Docker-based agent deployment and restarts its containers after pulling the current application image. |
| Blades CLI | no | 819 | A CLI agent keeps memory in plain files, loads Markdown skills and runs recurring tasks through a persistent scheduler. |
README in short
The platform separates CLI, client SDK and server components. It supports MCP connections and can expose agent tools through the protocol.