Does Nestor compete with Claude Code?

No. Claude Code is Anthropic's official CLI for Claude. Nestor is a multi-LLM agent platform with audit log, RBAC, and missions. They compose: Claude Code as the human-in-terminal seat, Nestor as the MCP server providing tools and missions.

Can Claude Code call Nestor?

Yes, via MCP. Add nestor-sh as an MCP server in your Claude Code mcp.json config. Claude Code then sees 15 nestor_* tools (run mission, query audit log, search knowledge graph, etc.).

Do I need an Anthropic API key for Nestor?

Optional. Nestor supports 7 providers: Anthropic, OpenAI, Gemini, Grok, Mistral, Ollama (local, free), and OpenRouter (300+ models). Use any combination.

Nestor vs Claude Code — when to use each (and how to use both)

TL;DR

Claude Code is Anthropic's official CLI. Tight Claude integration, free for Pro/Max subscribers, deep filesystem + shell tools, MCP client built-in.

Nestor is multi-LLM (Claude + 6 others), BYOK, ships with audit log, missions, DAG, RBAC. It's an MCP server that Claude Code can plug into.

This isn't competitive. The hybrid setup is the whole pitch — Claude Code as the human-driving-it-from-terminal seat, Nestor as the agent server with audit trail. Run both.

Feature matrix

Updated 2026-04-26 against Claude Code 1.0.x and Nestor 3.5.0.

Feature	Claude Code	Nestor
License	Proprietary (Anthropic)	MIT
Distribution	`@anthropic-ai/claude-code`	`nestor-sh`
Primary surface	CLI	CLI + Studio + MCP server
Multi-LLM	Claude only	7 providers
BYOK	Optional, defaults to subscription	Always
Free Claude Pro/Max integration	Yes (free for subscribers)	N/A — BYOK only
MCP client	Yes	Yes
MCP server	No	Yes — 15 nestor_* tools
IDE integration	VS Code / JetBrains plugin	Studio web UI
Audit log	No	Hash-chained
RBAC / multi-tenant	No	Yes
Mission orchestrator	No	Yes
DAG workflows	No	Yes
Cron schedule	No	Yes
Approval engine	Per-tool prompt	off/smart/manual + HMAC tokens
Budget caps	No	per-run + per-day
Persona / agents	CLAUDE.md	BMAD+ stack (5 personas)
Sandbox	Permission prompts	Rust core + optional Docker
On-prem	No (talks to Anthropic)	Yes, fully

When to pick which

Pick Claude Code if…

You're a solo dev with Claude Pro/Max who wants the best Claude UX
You want zero setup — it just works
You're happy with Claude as your only model
You don't need audit log, RBAC, or multi-tenant
You want Anthropic's officially blessed tool
You want the deepest IDE integration via the official plugin

Pick Nestor if…

You need multi-LLM with hot-swap (Claude + 6 others)
You need hash-chained audit log for compliance
You're running unattended missions on a server
You need RBAC + multi-tenant for a team
You want on-prem or air-gapped deployment
You want an MCP server to expose tools to other AI clients

The hybrid setup — the whole pitch

Run Nestor as an MCP server, plug Claude Code into it. Claude Code keeps its inline UX, Nestor adds missions + multi-LLM + audit + RBAC.

Step 1 — start Nestor MCP server:

npx nestor-sh mcp serve

Step 2 — add to Claude Code's ~/.claude/mcp.json:

{
  "mcpServers": {
    "nestor": {
      "command": "npx",
      "args": ["-y", "nestor-sh", "mcp", "serve"],
      "env": { "NESTOR_API_TOKEN": "${NESTOR_API_TOKEN}" }
    }
  }
}

Claude Code now sees 15 nestor_* tools: nestor_run_mission, nestor_skill_invoke, nestor_audit_query, nestor_search_kg, etc. You can call them from inside Claude Code's chat.

Use case: tell Claude Code "run an OSINT mission on acme.com via Nestor". Claude Code calls nestor_run_mission. Nestor executes server-side with mission DAG + audit log + budget caps. Returns structured result. Claude Code shows it inline.

This is the most powerful agent stack you can run today, and it's all OSS + BYOK.

What Nestor adds that Claude Code doesn't

Multi-LLM hot-swap: Claude rate-limits → fallback to GPT-4o or Gemini Flash mid-mission, no state loss
Mission orchestrator: planner → executor → evaluator → reporter, structured outputs, resumable
DAG workflows: cron-scheduled multi-step pipelines
Hash-chained audit log: verifyImmutableAuditChain(), tamper-evident, SOC 2 friendly
RBAC + multi-tenant: team setup with per-user permissions
Budget caps: per-run + per-day USD ceilings, hard cap toggle
MCP server: 15 nestor_* tools exposed to Claude Code, Cursor, Cline, Continue
Studio dashboard: 30+ pages for missions, audit, agents, scheduler, costs

What Claude Code does better than Nestor

Deep IDE integration: VS Code / JetBrains official plugins
Anthropic blessing: official tool, evolves alongside Claude API
Free tier: included with Claude Pro/Max — Nestor is BYOK API only
Setup time: zero config for Claude users vs Nestor's interactive wizard
Larger ecosystem: more Anthropic-blessed extensions and integrations
Polished UX: Anthropic's product team has resources we don't

Honest verdict

Claude Code is the best way to use Claude. If Claude is your primary model and you live in a terminal, install it. We don't try to compete on that.

Nestor isn't trying to replace it. We're the MCP server it can call when it needs multi-LLM, audit, or autonomous server-side missions. Different shape, complementary purpose.

Run both. The hybrid setup is the most powerful agent stack on the market today.

Try Nestor

npx nestor-sh install
npx nestor-sh mcp serve   # MCP server for Claude Code

Or browse the quickstart, the MCP integration guide, or the full changelog.