Clausona Account Manager

Switch Claude Code and Codex CLI accounts with one command while sharing each tool’s setup.

Stars
★ 50
Last updated
today
License
MIT
Primary language
TypeScript

At a glance

How it runs
CLI
Works with
Platform-specificCodex · Claude CodeClaude API (Partial support)
Setup effort
Low · running in minutes
You'll need
Node.js >= 20Claude Code CLI or OpenAI Codex CLIPowerShell 5.1+ on Windows, or zsh or bash on macOS/LinuxShell / CLINetwork accessLocal filesystemMCP Server
Typical use
A developer who uses personal and work Claude Code accounts on one machine and wants separate sign-ins and conversation histories.
Not a fit if
  • Users who want API profiles for Claude Code only and do not use a signed-in Claude Code account
  • Teams that need Codex API endpoint profiles
Source review
74/100 · Some gaps

What does this agent do, and when should you use it?

Clausona is a local profile manager for Claude Code and OpenAI Codex CLI, exposed through the `clausona` command, its `csn` alias, and an interactive terminal dashboard. Each account has its own config directory for sign-in and conversation history; an installed shell hook sets the relevant environment variable when `claude` or `codex` starts, then runs the original CLI. Selected settings, plugins, and skills are linked between profiles for the same tool, and session history can optionally be shared. The CLI and dashboard show each account’s 5-hour and weekly plan usage, with local Claude Code cost and token tracking and profile health checks also available. It runs on the user’s macOS, Linux, or Windows machine without a persistent service; quota lookups and update checks use the network.

Clausona requires Node.js >= 20 and at least one of Claude Code CLI or OpenAI Codex CLI. Run clausona init to discover and register already signed-in accounts; then use clausona add to create profiles, clausona use <profile> to select one, or clausona run <profile> to start a profile without changing the selection. The installer’s shell hook sets CLAUDE_CONFIG_DIR or CODEX_HOME for each claude or codex launch and executes the unmodified CLI. Clausona links selected configuration across profiles for each tool and keeps or merges conversation history according to the chosen session mode. clausona list and the dashboard read the tools’ own usage endpoints to show plan quota; clausona doctor checks profile health and clausona repair fixes broken shared links.

  1. A developer who uses personal and work Claude Code accounts on one machine and wants separate sign-ins and conversation histories.
  2. A developer switching among OpenAI Codex CLI accounts who wants to avoid manually juggling CODEX_HOME.
  3. A developer or team member who wants profiles for the same tool to share plugins, skills, hooks, and settings.
  4. A user who wants to compare remaining Claude or Codex plan quota across accounts before choosing one.
  5. A developer who needs to run sessions under different accounts in separate terminals using clausona run.
  6. A Claude Code user who wants a profile backed by an Anthropic API, gateway, or self-hosted model endpoint.

How do you install or deploy this agent?

Requires Node.js >= 20 and at least one of Claude Code CLI or OpenAI Codex CLI. macOS/Linux support zsh or bash; Windows requires PowerShell 5.1 or later. Use the installer for your platform:

curl -fsSL https://github.com/larcane97/clausona/releases/latest/download/install.sh | bash
irm https://github.com/larcane97/clausona/releases/latest/download/install.ps1 | iex

If neither CLI has been signed in, first run claude login or codex login. If Node is managed by a version manager, activate it in the current shell before running the macOS/Linux installer.

How do you use this agent?

Initialize and register existing accounts:

clausona init

Switch profiles, inspect accounts and quota, or open the interactive dashboard:

clausona use work
clausona list
clausona

Add and switch to another Codex account:

clausona add codex:work
clausona use codex:work

To create a Claude Code API profile, you need an existing Claude login and an endpoint. The key can be entered at the hidden prompt or supplied through standard input:

clausona add claude:gw --api --base-url https://openrouter.ai/api
clausona use claude:gw

What are this agent's strengths and limitations?

Pros
  • Manages Claude Code and Codex CLI accounts while keeping their sign-in credentials separate.
  • Shares linked plugins, skills, hooks, and settings to reduce repeated setup across profiles.
  • Shows 5-hour and weekly plan quota per account, with --refresh and --no-quota options.
  • Can run a profile without changing the selected account and includes a terminal dashboard and health checks.
  • Claude Code API profiles support Anthropic API, gateway, and self-hosted model endpoints.
Limitations
  • Requires Node.js >= 20 and at least one installed Claude Code CLI or OpenAI Codex CLI.
  • API profiles currently work only with Claude Code, and creating one still requires a Claude login.
  • MCP servers added with Claude’s claude mcp add command remain in each account’s config and are not shared.
  • Shared Codex conversation history is supported only on macOS and Linux; on Windows, another account’s thread must be resumed by ID.
  • Quota display and update checks need network access; the README does not state the cost of the CLI accounts or subscriptions.

How does this agent compare with similar options?

The README compares Clausona with claude-swap, codex-auth, aisw, and claude-code-profiles. Its listed distinctions include support for both Claude Code and Codex CLI, keeping history per account while sharing plugins and settings, and showing plan quota for every account.

Key facts side by side with the most closely related agents.

Agent Source review Form / cost Stars Updated Language Full support on
Clausona Account Manager This agent 74 · Some gaps CLI ★ 50 today TypeScript Codex · Claude Code
Herdr Terminal Agent Runtime 49 · Major gaps CLIFree ★ 43k 1d ago Rust —
Jean Desktop AI Development Assistant 47 · Major gaps Desktop appFree + model costs ★ 1.3k today TypeScript Codex · Claude Code · OpenAI API · Claude API
AGF Local Coding Session Finder 96 · Excellent CLIFree ★ 157 8d ago Rust Codex · Claude Code

How does FollowAgents rate this agent?

FollowAgents source review · FARS-2.1
Some gaps
74/ 100 5-point scale 3.7 / 5
Trust 18/29
Reliability 9/14
Adaptability 15/18
Convention 15/18
Effectiveness 12/13
Verifiability 5/8
Why each dimension lost points
Trust18 / 29 · 3.1/5

The README describes config directories, credentials, session history, shared files, quota lookups, and update checks. It also explains API-key input through a prompt, stdin, Keychain/local storage, or environment/command sources, supporting full marks for data-flow transparency. Switching is explicit, updates ask by default, removal preserves profile data, and session separation and repair are documented, supporting moderate scores for privilege, confirmation, and recovery. The dependency list gives versions but the supplied material shows no vulnerability review or mitigation. Publisher identity is unknown, and the repository gives a contribution path without establishing maintainer identity or update responsibility, so those scores remain limited.

Reliability9 / 14 · 3.2/5

The README gives broadly consistent descriptions of account separation, shared configuration, platform differences, and network behavior. The CI file lists linting, type checking, tests, builds, and bundle smoke checks; the Superset plugin test also covers errors, token redaction, and deletion safeguards. These are repository artifacts, not checks run for this review. Some reliability claims remain primarily documented assertions, and the supplied files omit core implementation, so consistency and dependency availability do not receive full marks. Error cases are illustrated, but the evidence does not establish clear messages for every path.

Adaptability15 / 18 · 4.2/5

The README addresses people using multiple Claude Code or Codex accounts on one machine and covers personal, work, organization, API endpoint, parallel-session, Superset, and quota scenarios. It clearly limits API profiles to Claude Code and describes Windows link behavior and session-sharing differences, earning strong boundary coverage. Invocation is mostly explicit through commands and a shell hook, though first-session API-key confirmation and environment variables that cause the hook to step aside show some trigger complexity. Platform and runtime requirements are detailed, but the core implementation is not supplied for compatibility verification.

Convention15 / 18 · 4.2/5

The README is organized into installation, quick start, features, operation, commands, FAQ, and limitations, with practical examples and platform-specific instructions. The MIT license text is complete. The package version is 0.5.1-beta and CI lists cross-platform checks, but no changelog is supplied, leaving version history thin. A contribution path is present, but maintainer responsibility and verified publisher identity are not established, so maintenance responsibility is only partly clear.

Effectiveness12 / 13 · 4.6/5

The CLI, JSON listing, interactive dashboard, and usage view help users select accounts and inspect quota. Per-account sign-in separation with shared settings and optional history addresses concrete repeated setup work. The README also explains the Node.js requirement, shell hook setup, key handling, and Windows link limitations, making user costs visible. Initial setup and platform constraints still add effort, and the supplied material lacks core implementation or runtime evidence to confirm every claimed benefit.

Verifiability5 / 8 · 3.1/5

The README gives traceable descriptions of commands, data locations, credential sources, limitations, and network calls. CI and plugin tests provide a second kind of repository evidence for some error handling and safeguards. The supplied files omit the core product implementation, and the README's comparisons are described as based on other projects' documentation rather than independent verification. CI describes intended checks but does not establish that they were run or passed, so scores are moderate.

Risks and how to mitigate them
  • This assessment uses only the supplied static material. Core implementation files are absent, so the documented security and compatibility behavior cannot be checked item by item.
  • API profiles handle keys. Review the selected storage method and the README's note about Claude Code's first-session key confirmation.
  • On Windows, some shared files may remain per-account copies; review the documented platform limitations before relying on shared settings.
Evidence confidence: Low Reviewed Oct 09, 2026 Reviewed revision 085e7303d61c New commits since this review; the score may not cover them
See the full review method →

FAQ

Does Clausona itself have a stated price?
The README identifies an MIT license but does not state the costs of the Claude Code or Codex CLI accounts, subscriptions, or APIs needed to use them, so it does not establish the total cost.
Can I resume a conversation under another account?
Conversation history can be shared by choice. Claude Code profiles can merge history with clausona config <profile> --merge-sessions; Codex shared thread storage is supported on macOS and Linux, while Windows users resume another account’s thread by its ID.
Which settings carry over when I switch profiles?
Plugins, skills, and selected settings are shared between profiles for the same tool; sign-ins remain separate. MCP servers added with claude mcp add stay in each Claude profile’s config and are not shared.
Can I use an API endpoint profile with Codex?
No. The README says API profiles are currently for Claude Code only.
Are there Windows caveats for shared configuration?
Windows requires PowerShell 5.1 or later. Developer Mode enables symbolic links for shared files; without it, some files that tools save as a whole remain per-account copies, and clausona doctor reports this.
View on GitHub ↗ Install ↓

Compare agents like this one

The same FARS review applied across the shortlist this agent qualifies for.

Related agents