Claude Code

Synara connects to the Claude Code runtime installed on your machine. Claude owns the account, model access, tools, and authentication; Synara owns the task, working environment, transcript, approvals, review surfaces, and Git lifecycle around the session.

Install

Use Anthropic’s current setup instructions. A common package-manager installation is:

npm install -g @anthropic-ai/claude-code

Synara looks for the claude executable. It can also use a native or Homebrew installation when that executable is available to the desktop process.

Verify the installation:

claude --version

Do not use sudo npm install -g. Fix the package-manager permissions or use an official alternative instead.

Authenticate

Start Claude Code in a terminal:

claude

Complete the authentication path appropriate to your setup. Claude Code supports Claude account subscriptions, Anthropic Console access, and supported enterprise backends such as Amazon Bedrock or Google Vertex AI.

Where supported by your installed version, inspect authentication directly:

claude auth status

Synara does not copy or replace Claude credentials.

Verify

Before opening Synara, verify Claude Code from a fresh terminal in a test repository:

cd /path/to/test-repository
claude --version
claude

Ask a harmless read-only question, then exit. This checks the binary, authentication, repository access, and provider service independently of Synara.

Connect to Synara

Capabilities in Synara

Claude Code sessions can expose provider-specific behavior through Synara, including:

• Model and effort selection
• Plan and execution modes
• Tool calls, file changes, and command output
• Approval requests and user questions
• Resumable sessions
• Claude subagents and tracked tasks where supported
• Context-usage information where available
• Native steering of a running turn

Native steering means a follow-up can be delivered to supported live Claude sessions without first cancelling the active turn. The provider still decides when and how to act on it.

Claude capabilities change with the installed CLI and account. Treat the controls shown in your current Synara session as authoritative for that runtime.

Fable 5.1 controls and usage

Synara models Fable 5.1 with always-on thinking and Low through Max effort. It does not expose a separate thinking toggle, a fast-mode lane, or the legacy ultrathink prompt option. These controls describe Synara's current integration; model access still depends on the Claude account and runtime.

When Anthropic returns a model-scoped weekly allowance, Synara shows a separate Fable window in Claude usage directly after the general weekly window. A missing Fable row means the authenticated account response did not provide that limit; it does not by itself mean the model is unavailable.

Updating

Synara can recognize common npm, Homebrew, and native Claude installations. Depending on the installation source, update Claude Code through its package manager or:

claude update

Restart Synara after replacing the executable.

Automatic compaction and explicit windows

The compaction picker defaults to Auto (Claude Code). In this mode Synara leaves the window unpinned so Claude Code can resolve its own model, account, settings, and environment behavior. Choose 200k or 1M when you want an explicit override; returning to Auto clears that override. A displayed model context capacity and an automatic compaction threshold are different values.

Context usage comes from the SDK summary when available, and a live model change refreshes the effective context metadata. An explicit window selection does not grant an account additional model access.

Troubleshooting

Claude is not detected

1. Run claude --version from a fresh terminal.
2. Compare the resolved executable with the custom binary path in Synara.
3. Restart Synara after changing PATH or installing Claude.
4. Remove an obsolete custom path rather than pointing Synara at a wrapper that no longer exists.

Claude is detected but not authenticated

Run claude directly and complete the provider’s own login flow. For enterprise backends, verify the required cloud credentials and environment variables in the same shell environment.

Models or effort options are missing

Update Claude Code, confirm the account exposes the expected model, and refresh provider status. For Fable 5.1, start a new task after refreshing so the current model catalog and capability ladder are applied. Model availability is account- and runtime-dependent.

A session behaves differently from the Claude terminal

Compare the selected model, permission mode, project settings, local instructions, and executable path. A different Claude installation or configuration source can produce different behavior.

Usage history and context compaction

Synara 0.8.4 corrects repeated Claude response accounting and distinguishes current context usage from accumulated processed tokens. Profile Stats counts verified completed model totals without adding cache or thinking tokens twice, and avoids counting mirrored subagents a second time.

Older unverifiable processed totals are hidden rather than presented as accurate. Historical recovery is partial: missing raw results and already-purged aggregates cannot be reconstructed, so older charts can show less usage or incomplete history after the update. This is an accounting correction, not evidence that your subscription usage or bill decreased.

A compaction icon indicates active context compaction. Wait for compaction to finish before treating a quiet transcript as a stalled response. Pending questions and drafted answers have recovery handling; see Tasks and runtime.

Official documentation

• Set up Claude Code
• Claude Code CLI reference

Continue with Your first task or Provider handoffs.