Local Operator
An open-source AI agent that lives in your terminal and works on your machine
Plans, runs tools, browses, spawns subagents, and remembers — all from a fast terminal UI
The Local Operator TUI mid-task: streamed responses, expandable tool cards, and one-line receipts for everything the agent does.
Local Operator is a terminal-native AI agent: describe what you want done and it does the work on your machine, asking before anything writes or executes. It is MIT-licensed and built to be lived in — sessions persist and resume, context compacts itself before it overflows, and the agent can schedule its own follow-ups.
📚 Table of Contents
- ✨ Why Local Operator
- 🚀 Quickstart
- 🖥️ A Tour of the TUI
- 🔌 Providers
- 🧰 What the Agent Can Do
- ⚙️ Headless & Server Modes
- 📱 Phone Access (Mobile Relay)
- 📦 Installation Options
- 🔧 Configuration & Credentials
- 🌟 Radient Agent Hub
- 🔒 Safety Model
- 📝 Examples
- 👥 Contributing
- 🙏 Credits and Acknowledgements
- 📜 License
✨ Why Local Operator
- A real terminal UI, not a REPL. A full-screen Textual app with streamed responses, expandable tool cards, session resume, 20+ built-in themes with live preview, and a status line that tells you what the agent is doing and what it costs.
- Sign in with the account you already have. OAuth login for OpenAI (ChatGPT), Anthropic (Claude), Kimi, xAI, Z.AI, and Qwen — or bring an API key, or run entirely offline with Ollama.
- Approval-gated execution. Reads are automatic; writes and shell commands
ask first.
/approvals auto(or--yolo) opts out deliberately, per session or as a saved default. - An agent workforce, not just an agent. Fan work out to concurrent
subagents with tool-restricted roles (a
reviewerthat cannot edit what it reviews), author reusable agent profiles, and save whole teams — a manager plus a roster — you launch by name with/team. Peek at, steer, pause, or cancel any worker mid-flight. - A session you can leave and come back to. Transcripts persist,
/resumepicks up where you left off, and context compaction runs itself before the window fills, so long sessions don't fall off a cliff. - Skills, MCP, scheduled wakes, web search, a browser tool — the agent's toolbox is broad, and everything it does leaves a visible receipt in the transcript.
🚀 Quickstart
Requires Python 3.12+.
pip install local-operator # pipx install local-operator on Linux (PEP 668)
The install provides both local-operator and its short alias lop — the
rest of this page uses lop. lop update upgrades that install from PyPI
and restarts the mobile daemon when the LaunchAgent is installed;
lop-update (hyphen) is the developer script that rebuilds the global
runtime from a local git checkout.
Sign in to a provider (or skip this — on an interactive terminal lop opens
in a setup state and walks you through /login; a headless or piped run prints
the exact commands to configure hosting, model, and a key):
lop login # lists login-capable providers
lop login anthropic # OAuth sign-in in your browser
lop login <provider> also sets it as your default hosting (and picks a
default model) when none is configured yet, so the very next lop just works.
Then start it:
lop
That's it. Type what you want done. esc stops the agent, /help lists
commands, /exit quits.
Prefer a fully local model? Install Ollama, pull a model, and point the agent at it:
lop --hosting ollama --model qwen2.5:14b
🖥️ A Tour of the TUI
Everything the agent does shows up as a card or a one-line receipt. Tool
cards expand (enter/space) to show the full command and output; the
status line tracks the current step, token usage, and cost.
When a tool call needs your sign-off, the approval prompt shows exactly what is about to run before anything touches your system:
Switching models is a picker, not a config file — /model lists every model
your signed-in providers offer, with fuzzy filtering:
Ask for parallel work and the agent delegates: the subagent dock shows each
worker's status, spend, and progress live, and you can open any of them to
watch its transcript. (This shot also shows one of the 20+ built-in themes —
/theme previews them live as you arrow through the list.)
Coming back later is /resume — a picker over your recent sessions, each
with its title and age:
And /usage answers the question every agent user has: how much provider
quota is left, and what the account has spent (the status line tracks the
current session's cost live).
Slash commands
/help shows the full table in-app. The highlights:
| Command | What it does |
|---|---|
/model |
Switch model for this session; /model default saves it for new ones |
/effort |
Show or set reasoning effort (shift+tab cycles) |
/approvals |
Set whether tools ask first (ask/auto; add default to keep it) |
/resume |
Pick a past conversation and continue it |
/new, /clear, /reload |
Fresh conversation · wipe the screen · relaunch this conversation on the current install |
/update |
Install the latest version from PyPI and relaunch |
/goal, /loop |
Set an objective, then iterate autonomously toward it |
/btw |
Ask a side question off the record — it never joins the conversation |
/compact |
Compact the context now (it also happens automatically) |
/usage, /context |
Provider quota and account spend · what's occupying the context window |
/provider, /login, /logout, /accounts, /credential |
Manage providers and stored credentials |
/search |
Configure web-search providers and load balancing |
/team |
Launch a saved team: /team <name> <request> puts a manager and roster on it |
/skills, /mcp |
List loaded skills · MCP servers |
/theme, /rename |
Pick from 20+ built-in themes (arrows preview live) · rename the session |
Keys worth knowing
- Type while the agent works — your message is delivered at the next step as steering, no need to wait.
esc— stop the agent without ending the session.ctrl+b— open an aside (side question) without losing what you were typing;ctrl+fpromotes the aside into the conversation.shift+tab— cycle reasoning effort.ctrl+l— clear the transcript (history is untouched).
🔌 Providers
One agent, your choice of brain. OAuth providers sign in through the browser and use your existing subscription; API-key providers prompt once and store the key locally; Ollama runs models on your own hardware.
| Provider | Access |
|---|---|
| OpenAI / ChatGPT | OAuth (browser or device code) or OPENAI_API_KEY |
| Anthropic / Claude | OAuth or ANTHROPIC_API_KEY |
| Kimi (Moonshot) | OAuth or KIMI_API_KEY |
| xAI / Grok | OAuth or XAI_API_KEY |
| Z.AI (GLM) | OAuth or ZAI_API_KEY |
| Qwen (Alibaba) | OAuth (token plan) or API key |
| Google Gemini | GOOGLE_AI_STUDIO_API_KEY |
| DeepSeek | DEEPSEEK_API_KEY |
| Mistral | MISTRAL_API_KEY |
| OpenRouter | OPENROUTER_API_KEY — one key, many models |
| Radient | RADIENT_API_KEY — automatic per-step model selection |
| Ollama | Local, no key, no network |
lop login # list login-capable providers
lop login openai # OAuth flow
lop login-status # what's signed in
lop logout kimi
Legacy --hosting <name> --model <name> flags keep working, and API keys can
be stored with lop credential update <KEY_NAME> (a masked prompt).
🧰 What the Agent Can Do
The agent's built-in tools, each with its own card in the transcript:
- Run things —
bash(shell commands),eval(a persistent Python kernel: variables survive across calls). - Work with files —
read,write,edit(surgical search/replace),glob,grep, pluslspfor Jedi-backed Python code intelligence. - Reach the web — load-balanced
web_searchacross seven providers and abrowsertool for pages that need rendering or interaction. - Stay organized — a visible
todolist for multi-step work,askto put real decisions back to you as a picker instead of a wall of text. - Work in the background —
taskspawns subagents,jobs/wait/hubmanage and talk to them, andwakeschedules future follow-ups ("check the build again in 30 minutes").
🔎 Web search
Search works out of the box — DuckDuckGo and Tavily's keyless endpoint are enabled by default, and requests rotate across providers with automatic fallback when one is rate-limited or down:
lop search list
lop search test "Python 3.13 release notes"
lop search enable perplexity
lop search setup brave --api-key
lop search setup tavily --oauth # official Tavily MCP server
lop search setup searxng --endpoint https://search.example.com
| Provider | Access | Default |
|---|---|---|
| DuckDuckGo | Credential-free | Enabled |
| Tavily | Keyless, TAVILY_API_KEY, or OAuth MCP |
Enabled |
| Perplexity | Anonymous or PERPLEXITY_API_KEY |
Disabled |
| Brave | BRAVE_API_KEY |
Disabled |
| Exa | EXA_API_KEY |
Disabled |
| SerpApi | SERPAPI_API_KEY |
Disabled |
| SearXNG | Self-hosted endpoint URL | Disabled |
The same controls are available in-app via /search.
🤝 Subagents, agent profiles, and teams
This is where Local Operator stops being a chatbot and starts being a staff.
Subagents. Ask for parallel work and the agent fans it out into concurrent background workers, then keeps working while they run. Each worker is addressable: peek at its transcript, send it a note, ask it a question, steer it onto a different course, pause it, or resume it later — all without burning its attention on status meetings.
Roles are capability boundaries, not just prompts. A subagent launched as
reviewer carries vetted review guidance and loses the tools to edit code
— it can read and run tests but cannot alter what it reviews, which is what
keeps a review honest. Packaged starters ship for reviewer, coder,
architect, manager, designer, and scout, and you can author your own
agent profiles: reusable roles and named specialists with their own
instruction sets, matched to tasks by semantic routing. When a profile gives
bad guidance, you fix the profile once — not every prompt that uses it.
Teams. A saved roster — a manager plus members with counts — layered with
two briefs the individual agents never hard-code: a collaboration brief (how
this group works together, who blocks a release) and a project brief (what
product this instance owns). Swap the project brief and the same roster staffs
a different product. /team lists your saved teams:
The team picker. Launch one with /team <name> <request> — the current agent becomes that roster's manager and delegates from there:
Sending a request to a team is one line — the manager breaks it down and puts the right roles on it.
Agents can also be managed from the CLI:
lop agents create "My Agent"
lop agents list
lop teams list
🧠 Skills and guides
Drop a SKILL.md (with optional reference files) into
~/.local-operator/skills/<name>/ and the agent indexes it semantically —
only the skills relevant to the current turn are surfaced, and their bodies
load on demand via skill://<name> reads, so your context isn't taxed by
knowledge you aren't using. /skills lists what's loaded.
🔗 MCP servers
Local Operator speaks MCP over the official SDK, with lazy tool loading: servers advertise a bounded summary, and individual tool schemas enter the context only when the agent actually enables them.
lop mcp add linear --url https://mcp.linear.app/mcp --oauth
lop mcp login linear # complete the OAuth flow
lop mcp list
Server configs are discovered from the project (.local-operator/mcp.json,
.mcp.json), your home directory, and best-effort imports of Claude Code,
Cursor, and VS Code configs — so servers you already configured
elsewhere just show up. See docs/mcp.md for the trust model
before enabling project-supplied servers.
⚙️ Headless & Server Modes
One-shot execution for scripts and automation:
lop exec "summarize the failures in ./test.log"
lop exec "long migration" --background # detach with a log file
lop exec "audit deps" --json # one JSON line per event
Exit code 0 on success — pipeline-friendly.
Server mode exposes the agent as a FastAPI service (used by the optional desktop UI):
pip install "local-operator[server]"
lop serve # http://localhost:1111, docs at /docs
Phone access — an optional session daemon lets you watch and steer your sessions from your phone. See Phone Access (Mobile Relay) below.
📱 Phone Access (Mobile Relay)
lop mobile turns the machine you run agents on into a phone-facing control
plane for every lop session on it. A single supervised session daemon
owns the web surface, and every interactive TUI session registers with it
automatically over an authenticated loopback socket. From your phone you can
watch transcripts stream, steer a running turn, switch model and effort, run
slash commands, drill into subagents, and start brand-new sessions. Sessions
you start from the phone also answer their own approval and ask prompts there;
for a terminal session those prompts are still answered at the terminal (the
phone shows that it is waiting).
A live lop session driven from a phone: the same transcript, tool cards, and composer as the TUI, mobilized.
You can ask Local Operator to set this up for you. Tell your agent something like "set up phone access" and it will walk through the install, confirm the health check and the closed auth gate, and get you the portal password through a channel you choose (or leave it in the Keychain for you to retrieve). If you would rather do it by hand:
lop mobile install # generate/keep the portal password, install the daemon, verify health
lop mobile status # install state, health probe, and registered sessions
lop mobile password # show or rotate the portal password
lop mobile logs -f # follow the daemon log
Once the daemon is up, every interactive lop you start publishes itself and
shows up in the phone list live. No extra flag per session.
Additional setup is required for remote access
The daemon binds loopback only (127.0.0.1:4098) and never a wider
address. That keeps it private by construction, so reaching it from your
phone over the internet needs a secure path you put in front of it, together
with an identity proxy so only you can open it.
The recommended method is a Cloudflare Tunnel with Cloudflare Access in front: the tunnel gives the daemon a public hostname without opening a port on your machine, and Access enforces sign-in before any request reaches loopback. A WireGuard-based mesh such as Tailscale is a good alternative if you would rather keep everything on a private network. Either way, do not change the bind address to expose the daemon directly; put the tunnel and the identity proxy in front of the loopback listener instead.
The portal itself is protected by a single password (Keychain-backed on
macOS, or LOP_MOBILE_PASSWORD for containers), and session cookies are
derived from it, so rotating the password invalidates every logged-in phone.
📦 Installation Options
The default install is deliberately small. Optional features live behind extras:
| Extra | Adds |
|---|---|
server |
The HTTP API server (lop serve) and background scheduler |
mcp |
Model Context Protocol client support |
images |
HEIC/HEIF image attachment decoding |
tokenizer |
Exact BPE token counting (estimated otherwise) |
lsp |
Jedi-backed symbol-aware Python navigation for the lsp tool |
all |
Everything above except lsp |
pip install "local-operator[all]" # quote it — shells glob the brackets
If you hit a feature whose extra is missing, the agent tells you which one to install instead of failing with an import error.
Nix: nix develop drops you into a reproducible dev shell via the
provided flake.nix.
Docker: docker compose up -d with the provided compose file.
🔧 Configuration & Credentials
Configuration lives at ~/.local-operator/config.yml:
lop config create # scaffold it
lop config list # every option, with descriptions
lop config edit <key> <value>
lop config open # open it in your editor
Commonly set values: hosting and model_name (skip the CLI flags),
conversation_length / detail_length (history kept verbatim vs
summarized), and tui.theme (any registered theme name — easier to set with
/theme, which previews live).
Credentials are stored in ~/.local-operator/credentials.env and never
echoed:
lop credential update TAVILY_API_KEY
lop credential delete TAVILY_API_KEY
OAuth tokens from lop login are stored separately and refresh
themselves.
🌟 Radient Agent Hub
Radient adds two optional capabilities:
- Automatic model selection —
lop --hosting radientpicks the best model per step to balance quality and cost, no--modelneeded. - Agent sharing — push your agents to the public hub, pull agents others published:
lop credential update RADIENT_API_KEY
lop agents push --name "My Agent"
lop agents pull --id "<agent_id>" # no key needed to pull
🔒 Safety Model
- Approval tiers. Read-only tools run automatically; anything that writes
files or executes commands prompts first, showing the exact command.
/approvals autoor--yolodisables prompts only when you say so. - Visible receipts. Every tool call leaves a card or one-line receipt in the transcript — there is no invisible action.
- Local-first options. Run Ollama models for closed-circuit operation where nothing leaves your machine.
- MCP trust model. Project-supplied MCP configs are treated as trusted input and warned about on first connect — see docs/mcp.md.
- Credential hygiene. Keys live in a local credential store, are entered through hidden prompts, and are kept out of transcripts.
📝 Examples
👉 The example notebooks show real tasks completed with Local Operator, saved from live sessions:
- 🔄 Automated commit message generation from git diffs
- 🔀 End-to-end pull request automation — creation, review, template completion
- 🔢 MNIST digit recognition — 99.3% accuracy on the Kaggle competition
- 🏠 House price prediction with XGBoost — top 5% Kaggle score
- 🚢 Titanic survival prediction with LightGBM
- 🌐 Web research and data extraction — scraping a sanctions list
- 📈 Business pricing analysis — optimal subscription pricing
👥 Contributing
We welcome contributions! See CONTRIBUTING.md for how to
submit bug reports and feature requests, set up a development environment,
and open pull requests. docs/ covers the architecture
(REWRITE.md), benchmarks, and verification evidence.
🙏 Credits and Acknowledgements
Local Operator stands on the shoulders of the broader open-source agent community. Several aspects of this harness's implementation were shaped by studying and drawing inspiration from the projects below. We're grateful to their authors and contributors for building in the open.
- opencode — created by
Dax Raad (
thdxr) and the Anomaly (formerly SST) team. Its terminal-native, model-agnostic coding-agent design informed our thinking on the interactive CLI experience and provider-agnostic model handling. - oh-my-pi — authored and maintained
by Can Bölük (
can1357), building on Pi by Mario Zechner (mariozechner). Its approach to agent orchestration and harness ergonomics inspired aspects of our subagent and tooling implementation.
Inspiration drawn from these projects informed our own independent implementation; any mistakes here are our own.
A note on reuse and credit
All of the projects above are MIT-licensed, as is Local Operator itself. Under the MIT license you are free to draw inspiration from or reuse code from Local Operator in your own work. Verbatim reuse of the code requires retaining the copyright notice and license text, as the license states. Beyond that legal minimum, we simply appreciate credit where credit is due — an acknowledgement of the projects and people whose work you build on, in the same spirit as the credits above. It costs little and it keeps open source healthy.
Core contributor: Damian Tran <damian@gominerva.com>.
📜 License
MIT — see LICENSE for details.
1 maintainer
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Filter files by name, interpreter, ABI, and platform.
If you're not sure about the file name format, learn more about wheel file names.
Copy a direct link to the current filters
File details
Details for the file local_operator-0.28.1.tar.gz.
File metadata
- Download URL: local_operator-0.28.1.tar.gz
- Upload date:
- Size: 2.4 MB
- Tags: Source
- Uploaded using Trusted Publishing? Yes
- Uploaded via:
twine/7.0.0 CPython/3.13.14
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
2d31499d2bc303660560c9343e3402222e7a607fe947fc474be4a6f6d2723dc4
|
|
| MD5 |
6a1c04586b8122dfd24e794ca80c444b
|
|
| BLAKE2b-256 |
e04a2c27dc63512ce69e82194db44808aa435b807548bf50880a8dcfbe2056eb
|
Provenance
The following attestation bundles were made for local_operator-0.28.1.tar.gz:
Publisher:
publish.yml on damianvtran/local-operator
-
Statement:
-
Statement type:
https://in-toto.io/Statement/v1 -
Predicate type:
https://docs.pypi.org/attestations/publish/v1 -
Subject name:
local_operator-0.28.1.tar.gz -
Subject digest:
2d31499d2bc303660560c9343e3402222e7a607fe947fc474be4a6f6d2723dc4 - Sigstore transparency entry: 2579639273
- Sigstore integration time:
-
Permalink:
damianvtran/local-operator@ca98ddb900ad419c90ed90d3a1dfe978512c9b30 -
Branch / Tag:
refs/tags/v0.28.1 - Owner: https://github.com/damianvtran
-
Access:
public
-
Token Issuer:
https://token.actions.githubusercontent.com -
Runner Environment:
github-hosted -
Publication workflow:
publish.yml@ca98ddb900ad419c90ed90d3a1dfe978512c9b30 -
Trigger Event:
release
-
Statement type:
File details
Details for the file local_operator-0.28.1-py3-none-any.whl.
File metadata
- Download URL: local_operator-0.28.1-py3-none-any.whl
- Upload date:
- Size: 2.5 MB
- Tags: Python 3
- Uploaded using Trusted Publishing? Yes
- Uploaded via:
twine/7.0.0 CPython/3.13.14
File hashes
| Algorithm | Hash digest | |
|---|---|---|
| SHA256 |
ba1379e7191342797ac1a1afc2c9a9cbc463125a4881a20fd8a274e11c5d1cb7
|
|
| MD5 |
e747c022cedadcd8206dce9d90143fbb
|
|
| BLAKE2b-256 |
cdb92f2c416788215f48eba14c099afc91fa0f84a8d2c9eb750fb923d00c8228
|
Provenance
The following attestation bundles were made for local_operator-0.28.1-py3-none-any.whl:
Publisher:
publish.yml on damianvtran/local-operator
-
Statement:
-
Statement type:
https://in-toto.io/Statement/v1 -
Predicate type:
https://docs.pypi.org/attestations/publish/v1 -
Subject name:
local_operator-0.28.1-py3-none-any.whl -
Subject digest:
ba1379e7191342797ac1a1afc2c9a9cbc463125a4881a20fd8a274e11c5d1cb7 - Sigstore transparency entry: 2579639276
- Sigstore integration time:
-
Permalink:
damianvtran/local-operator@ca98ddb900ad419c90ed90d3a1dfe978512c9b30 -
Branch / Tag:
refs/tags/v0.28.1 - Owner: https://github.com/damianvtran
-
Access:
public
-
Token Issuer:
https://token.actions.githubusercontent.com -
Runner Environment:
github-hosted -
Publication workflow:
publish.yml@ca98ddb900ad419c90ed90d3a1dfe978512c9b30 -
Trigger Event:
release
-
Statement type:
