Valo — Desktop App for Claude Code
The first context manager for Claude.
Claude Code builds software from conversation. Valo gives it a workspace — context control, live visualization, security, and a UI anyone can use. No terminal required.
One click. No spam. Just an update when we launch.
Without Valo
Lose context when you run out. Hit stop and pray your session survives. Install a package and hope it is not compromised.
With Valo
Trim context and keep going. Pause Claude safely — no data loss. Untrusted packages get blocked before they reach your machine.
Context management and codebase visualization — built in
Live Map
Your codebase as a living graph. Nodes light up as Claude works. Trails show where it went and why.
Context ControlPRO
Claude gets sharper when it carries less. Hide tools and skills you don't need — fewer tokens, clearer thinking. When a conversation gets long, hand off — faster than compaction, and nothing gets lost. A new chat opens automatically with full context, and Claude picks up right where it left off. The CLI has about context limits and compaction failures.
Codebase IntelligencePRO
Valo studies your codebase and remembers what it finds — across sessions. Hub files, tags, and summaries get indexed so Claude can search instead of re-reading.
Your Work Survives
Stop Claude, close the lid, lose your connection — your chat is still there when you come back. Valo repairs broken session files automatically. The CLI has about lost sessions and about a stop button that corrupts your chat. Valo's pause button, session repair, and duplicate deduplication route around all of it.
Claude Code security — secrets never reach the AI
Other tools sandbox the process. Valo sandboxes the secrets. The CLI has about exposed credentials and about files deleted without consent.
KeyguardPRO
Secrets live in an encrypted vault that travels with your project — AES-encrypted, unlocked by your password, cross-platform. Injected into commands at runtime, scrubbed from output. API keys and tokens never enter the conversation — never reach Anthropic's servers.
Guards
Safety scripts that block dangerous commands, automate workflows, and enforce rules. Four ship built-in. Add your own. Claude cannot bypass them — they run outside the conversation.
Trusted Terminal
Claude can install packages and configure your Mac. When admin access is needed, macOS asks you directly — your password never leaves your machine.
Streamer Mode
Redacts 30+ secret patterns on screen. Custom blocklist for extra words and phrases.
Excluded PathsPRO
Block specific folders from Claude's access. Keep sensitive directories off-limits.
Untrusted Install ProtectionPRO
New packages are blocked until they have been published long enough to trust. One toggle stops supply chain attacks, typosquatting, and malicious freshly-published packages before they reach your machine.
Per-Tool Control
Granular toggles for every tool Claude can use. Allow reading freely but require approval for writes. Fine-grained control over what Claude can do without asking.
Custom AI agents — no code, no framework
An agent is just a description of what you want done. Create skills, define workflows, compose agents — all from conversation. No code, no framework. The CLI has about memory that doesn't persist — Valo's Active Memory fixes that.
Build Your Own Agents
Create specialized agents from simple markdown files. Describe what you want — Claude builds it. No code required.
Custom Skills
Slash commands you define. Type /your-skill and Claude follows your playbook. Share them across projects.
Hide From Context
Hide any tool or skill from Claude's context. Fewer tools loaded means fewer tokens spent — and Claude thinks more clearly, like working at a clean desk.
The Left Panel
Skills, agents, guards, and every tool Claude has — browsable and toggleable. See exactly what it can do, turn off what it should not.
Agent Transparency
Agent names shown live in the timeline. Always know which agent is working and what it is doing.
Active Memory
Hard-won knowledge resurfaces automatically. Concept pages with gotchas and invariants appear when Claude touches related files.
Access Claude Code from your phone
The CLI has requesting mobile access. Valo ships it today — text, voice, photos, permissions, all from Telegram.
On the Go (Telegram)
Talk to Claude from your phone. Text, voice, and photos. Thinking blocks, tool calls, and responses stream live. Approve permissions from anywhere.
ValoLinkPRO
Connect to a Claude on another machine across the internet. Share a 6-character code, collaborate in real time. Tools stay locked for safety.
Everything a Claude Code GUI needs
The CLI has about permission fatigue — clicking Allow hundreds of times per session. Valo's per-tool toggles fix that in one click.
Tuned ModePRO
Claude in Valo is different from Claude anywhere else. We tune it to take action — deploy your site, manage your backups, install what you need — without asking you to open a terminal. It stays focused longer, too.
MultiviewPRO
Run multiple chats side by side — 2-up, 2x2, or 3x2 grid. See everything at once.
Cross-ChatPRO
Claude in one project can ask Claude in another — with that project's full codebase and context. Persistent collaboration across tabs and projects.
Valo Browser
A browser built for Claude — browse, automate, and debug without running out of context.
Live Preview
Claude writes code, you see it render. React, HTML, Tailwind — live in the right panel.
Nerve MapPRO
Function-level call graph. Claude sees how everything connects — what calls what, blast radius, dead code.
VoicePRO
Talk to Claude. Local Whisper transcription — no cloud. It knows your codebase and spells your variable names right.
Forever History
Every chat lives forever. Across restarts, reboots, reinstalls — searchable, archivable, restorable.
Valo SearchPRO
Claude finds what it needs — 2x faster and more accurate than default tools. Your entire project, searchable by meaning. Hybrid retrieval: semantic embeddings, BM25, PageRank, RRF fusion, call-graph expansion.
Custom Data PathPRO
Store conversations anywhere — encrypted drives, iCloud, external storage. Your data, your location.
Scheduled Tasks
Claude wakes up on a schedule and picks up where it left off. Set it and forget it.
Claude Knows Valo
Claude has a full manual for the app. Every button, every setting — just ask. You don't need to learn the tool.
One-Click Permissions
One click approves permissions for the whole task. No popups, no interruptions — Claude just works.
Desktop Automation
Claude can control your desktop — click, type, scroll, take screenshots of any app window. Automate repetitive tasks across applications, not just code.
It's also fun
What does Valo connect to?
Valo (the app)
updates.getvalo.devApp updates
licensing.getvalo.devLicense validation (Pro)
No analytics. No telemetry. No tracking.
Claude CLI (underneath)
api.anthropic.comAI inference
claude.aiAuthentication
Anthropic also collects optional telemetry (feature flags, crash reports) — disable with CLAUDE_CODE_DISABLE_NONESSENTIAL_TRAFFIC=1
Don't trust us — verify yourself. Run a network monitor while using Valo.
Trust & transparency →