Comparison

Nairi vs Claude Tag

Anthropic's Claude Tag joins your Slack channels as a first-party teammate running on Opus 4.8. Nairi gives you the same multiplayer team agent in Slack or Discord, but the context (rules, skills, MCP connections, vault, memory) lives in a workspace you control, with the harness and model swappable per agent. Here's an honest side-by-side.

Try Nairi free Read our take on context lock-in

TL;DR

Claude Tag is Anthropic's Slack teammate (beta, launched June 23, 2026, gated to Claude Team and Enterprise). You tag @Claude in a channel, it joins as a multiplayer teammate with access to the channels, tools, data, and codebases you grant, and it runs on Opus 4.8 with ambient mode that can chime in and follow up on its own. The prior Claude in Slack app sunsets on August 3, 2026.

Nairi solves the same job from a different architectural angle. Same multiplayer shape (one shared team agent in a channel, in Slack and Discord), but the operating memory of your org (rules, skills, MCP connections, vault, channel permissions) lives in a workspace you control, not inside Anthropic. You also pick the harness per agent (Claude Code, Codex, OpenCode, Cursor) and the runtime is an open-source Go daemon you can self-host.

Honest framing: if your team is already on Claude Team or Enterprise, Slack is your only surface, and you're committed to Anthropic for the long haul, Claude Tag is the simpler choice. Pick Nairi if you want harness and model optionality, Discord support, a self-host option, or you care about owning the context layer separately from the model vendor.

Side by side

Architectural differences first, then the areas where Claude Tag wins.

Dimension

Nairi

Claude Tag

Where your context lives

Rules, skills, MCP connections, vault, channel permissions, and memory live in a workspace you control. Exportable. If you leave Nairi, you keep the operating memory of how your org works.

Context accumulates inside Anthropic's product: channel access, connected tools, learned history, ambient observations. It's the thing that makes the agent useful, and it's the thing that's hardest to move.

Model portability

Per-agent choice of provider key. Today Claude, tomorrow GPT, next month a local model. The agent's context follows the swap.

Claude Opus 4.8 on Anthropic's cloud. No GPT, no Gemini, no open-source model. The intelligence and the workflow surface ship from the same vendor.

Harness choice

Per-agent choice of Claude Code, Codex, OpenCode, or Cursor. Mix harnesses across the same org. Whichever one fits the job.

Anthropic's proprietary agent loop. The launch describes it as "the beginning of an evolution of Claude Code," but the harness is Anthropic's, not configurable.

Discord support

Slack and Discord are both first-class. The same agent runs in either, with the same rules, skills, and vault.

Slack only. The beta is gated to Slack channels for Claude Team and Enterprise customers. Discord, Telegram, and iMessage aren't part of this product.

Plan gate

Team subscription starting at $20/month. No requirement to be on a specific provider's Enterprise plan.

Claude Team or Claude Enterprise only. Pro and Max users are not eligible for the Claude Tag beta. The prior Claude in Slack app sunsets on August 3, 2026.

Run the agent on your own hardware

The nairid daemon is open-source Go. The full agent loop runs inside the daemon. Docker, Kubernetes, a VM, or an air-gapped box.

Anthropic-managed cloud only. There's no self-host option. Orgs with strict data-residency or air-gap requirements can't run Claude Tag at all.

Org-level customization

Per-org rules, per-agent rules, per-agent MCP configs, org-wide shared skill library, org-scoped vault. Set once at the org level, applies everywhere.

Admins can restrict tools and channels per Claude identity, set spending limits, and view activity logs. Customization happens through Anthropic's admin surface, not workspace-level config you can export.

Repo and tool sources

The daemon connects to any codebase you have credentials for (GitHub, GitLab, Bitbucket, self-hosted Git) and any MCP server you point it at.

GitHub plus the Anthropic-curated tool catalog. The selected channels, tools, data, and codebases are granted from within the Anthropic admin surface.

First-party model

Multi-vendor by design. No first-party model. Whatever ships next quarter, you swap to it with a config change.

Claude Opus 4.8 on day one. No compatibility risk on Claude version bumps. Whatever Anthropic ships next, it's there immediately.

Ambient mode

Scheduled jobs and event triggers (cron, webhooks) ship today. Free-form passive channel observation, where the agent decides on its own when to chime in, is not in the product.

Ambient mode out of the box. Claude can flag relevant info from across the org, follow up on unresolved threads, and schedule its own tasks across hours or days.

Frictionless install for existing Claude teams

Install the Nairi Slack or Discord app, pick channels, mention an agent. New vendor relationship, two minutes to set up.

If your team is already on Claude Team or Enterprise, Claude Tag is one click away. No new product to onboard. Introductory launch credits are included for eligible orgs.

Pricing model

Flat team subscription. No per-API-call billing. The team shares the agent and the quota.

No separate Slack fee. Bundled into the existing Claude Team or Enterprise plan. For teams already paying for Claude seats, there's no additional line item.

Claude Tag descriptions reflect Anthropic's announcement and public documentation as of June 2026. Claude Tag is in beta. Verify the current behavior at anthropic.com.

Rent the intelligence, own the context

The structural difference between Nairi and Claude Tag isn't about features. It's about where the part of the system that compounds, your team's operating memory, ends up living. Both products ship a multiplayer teammate you mention in a channel. The interesting question is what happens a year from now, once the agent has accumulated context.

Models move. Cheap last year, free next year, replaced the year after. Switching from Claude to GPT to Gemini to whatever ships next quarter is, for most teams, a config change. The intelligence is rentable. What isn't rentable is the slow-built memory of how your company actually works: the Slack scar tissue, the exception paths, the implicit owners, the "we tried that in Q2, here's why it failed" knowledge.

Claude Tag is a first-party product from a model vendor. The context, by design, accumulates inside Anthropic. That's not a flaw; it's the business model. If Anthropic stays the best model vendor for the next five years and you stay on Team or Enterprise, the bundling is a feature.

Nairi inverts that. Rules, skills, MCP connections, vault, channel permissions, and memory live in a workspace you control and can export. You point the agent at whichever model is best this month. If you decide to leave Nairi tomorrow, you take your context with you. The platform sells you the workflow surface; the intelligence is yours to rent from whoever's best.

Bring your harness. Bring your model.

Claude Tag runs on Claude Opus 4.8 on Anthropic's cloud, with Anthropic's proprietary agent loop. Anthropic positions it as "the beginning of an evolution of Claude Code," but the harness is theirs and the model is theirs. There's no Codex, no OpenCode, no Cursor, no GPT, no Gemini, no local model.

Nairi treats the harness as a per-agent decision. Run one agent on Claude Code for the writing-heavy work, one on Codex for the tasks where it has the edge, one on OpenCode for the open-source path, one on Cursor for the IDE-integrated workflow. Each agent has its own provider key. You can mix them in the same org and pick the right tool per job.

The bet here is structural. Anthropic ships an Opus 4.8 today and an Opus 4.9 next quarter, and the agent gets better on Anthropic's roadmap. Nairi customers get the same upgrade with one config change, plus the option to switch entirely if the best model in 2027 isn't Anthropic's. The harness layer is a config setting, not a vendor decision.

Same agent in Slack and Discord

Claude Tag works in Slack. That's it. The beta is gated to Claude Team and Enterprise customers on Slack. Discord isn't part of this product, and Anthropic's existing Channels feature for Discord, Telegram, and iMessage is a different architecture entirely (it pushes events into a local harness on your machine).

Discord is the workspace surface for a real slice of the team-AI-agent market. Dev tools companies, open-source maintainers, gaming teams, DAOs, internal community ops at consumer products. A team-AI-agent product that doesn't ship a Discord surface forecloses that market by design.

Nairi treats Slack and Discord symmetrically. Same agent, same vault, same rules, same skill library, same audit. Add the agent to a Slack channel, add it to a Discord channel, mention it from either. Internally the integration is the same. Teams that run both surfaces (engineering on Slack, community on Discord) can have one agent answer in both.

Run the agent on your own hardware

Claude Tag is Anthropic-cloud only. The agent runs in Anthropic-managed infrastructure and there's no path to move it onto your own. For most teams that's fine. For regulated industries, government contractors, financial services, healthcare, or any org with a real data-residency, IP allowlist, or air-gap requirement, it's a hard blocker.

Nairi's nairid daemon is open-source Go. The full agent loop runs inside the daemon: receiving messages from Slack or Discord, talking to the harness, executing tools, posting back. You can run it in Docker, on Kubernetes, on a single VM, on an air-gapped box, anywhere the binary fits. Most teams take the managed offering because they don't want to run a daemon. The self-hosted path exists for the teams that need it.

Honest framing: if your team has no compliance constraint and no preference for owning the runtime, this difference doesn't matter day to day. If you've ever had to pull a vendor's product out of procurement because the data-residency or self-host story didn't check out, you know how often this kind of thing decides the deal.

Where Claude Tag wins

Honest comparison pages are honest both ways. Here's where we wouldn't try to claim parity.

Ambient mode out of the box. Claude can chime in unprompted, flag things from across the org, and follow up on unresolved threads on its own. Nairi ships scheduled jobs and event triggers (cron, webhooks, handlers), but free-form passive channel observation isn't in the product today. If ambient behavior is the main job-to-be-done, Claude Tag has the edge.
Frictionless install if you're already on Claude. If every engineer on your team has a Claude Team or Enterprise seat and you're committed to Anthropic's roadmap, Claude Tag is one click away. No new product to onboard, no new vendor relationship, no provider key to manage.
First-party model and harness. No compatibility risk on Claude version bumps. Whatever Anthropic ships next, Claude Tag gets on day one. Today that means Opus 4.8. Anthropic reports 65% of their own product team's code is created by their internal version of the same product.
Bundled pricing. No separate Slack line item. Claude Tag rolls into the existing Claude Team or Enterprise plan, and eligible orgs get introductory launch credits. For teams already paying for Claude seats, the marginal cost of adoption is zero.
Admin surface for Claude identities. Admins can restrict tools and channels per Claude identity, set spending limits, and view activity logs. Same shape as Nairi's per-agent config, but built into Anthropic's admin console rather than a separate platform.

Common questions

What teams ask when they're weighing the two.

Claude Tag is Anthropic's Slack teammate, launched as a beta on June 23, 2026, for Claude Team and Claude Enterprise customers. You tag @Claude in a Slack channel, Claude joins as a multiplayer teammate, gets access to the channels, tools, data, and codebases you grant it, and you delegate work the way you would to a coworker. It runs on Claude Opus 4.8 on Anthropic's cloud. The prior Claude in Slack app sunsets on August 3, 2026.

Yes, with a clear architectural split. Both ship a multiplayer team agent you mention in a channel. The difference is where the operating memory of your team lives. Claude Tag is a first-party Anthropic product: same vendor for the model and the workflow surface. Nairi is multi-vendor: workspace-scoped rules, skills, MCP connections, and vault, with harness and model choice per agent. Same surface, different lock-in story.

Models are rentable. The slow-built memory of how your company actually works is not. The Slack scar tissue, the exception paths, the implicit owners, the unfinished threads, the "we tried that in Q2" knowledge. Once that memory accumulates inside a single vendor's product, moving it elsewhere is structurally hard. Nairi keeps that memory in a workspace you control and can export. Claude Tag keeps it inside Anthropic. Neither is wrong; the trade is who owns the part that compounds. Our full take is in the blog post on Claude Tag and context lock-in.

Yes. Most Nairi customers do. Claude Code is the default harness for many agent types, and Nairi uses your provider key to bill Claude usage directly to you. The difference from Claude Tag is that you can swap to a different harness or a different model later without rebuilding the agent's context.

Because the bet that one vendor stays best forever has historically not aged well. New models with new capability curves arrive every few months. Some workloads will fit Claude best, some Codex, some OpenCode, some Cursor, some a local model. Nairi makes the choice a config setting on an agent, not a structural choice baked into the platform. If Claude stays best for your team's workloads, you stay on Claude and lose nothing. If something better ships, you swap one setting and the agent's context comes with it.

Claude Tag does not extend to Discord. The beta is Slack-only and Claude Team or Enterprise gated. Nairi treats Slack and Discord as symmetric chat surfaces: the same agent, the same vault, the same rules, the same audit. If your team's primary surface is Discord (common for dev tools, open-source maintainers, gaming teams, and DAOs), Claude Tag isn't available and Nairi is the realistic pick.

Not in the free-form sense Claude Tag ships today. Nairi has scheduled jobs and event triggers (cron, webhooks, message handlers) for proactive behavior with explicit triggers. What we don't yet ship is passive channel observation where the agent decides on its own when to chime in across the org. If ambient mode is the main job-to-be-done and you're already on Claude Team or Enterprise, Claude Tag is the better pick for that specific workflow.

Claude Tag is Anthropic-cloud only. There's no path to move it onto your own infrastructure. Nairi's nairid daemon is open-source Go. You can run the full agent loop in Docker, on Kubernetes, on a single VM, or on an air-gapped box. If your team has data-residency, IP allowlist, or air-gap requirements, only the self-host path supports them.

Nairi has an org-scoped vault. Secrets are scoped per agent, audited at the org level, and injected at request time by a secret proxy. The agent's process environment never holds plaintext credentials. Claude Tag exposes tool and channel restrictions, spending limits, and activity logs through Anthropic's admin surface, but the secret-management primitive itself lives inside Anthropic's environment. Same shape, different blast radius if you ever decide to leave.

Claude Tag is available to Claude Team and Claude Enterprise customers only. Pro and Max users are not eligible for the beta. Eligible orgs receive introductory launch credits. Nairi starts at $20/month for a team subscription, no provider-plan requirement. You bring your own Anthropic, OpenAI, or other provider key for model usage.

Pick Claude Tag if all of these apply: your team is already on Claude Team or Enterprise; Slack is your only chat surface; you're committed to Anthropic's models and harness; ambient mode is core to the workflow; you don't need a self-host option; and you're comfortable with operating memory accumulating inside Anthropic. If those all hold, Claude Tag is the simpler choice and we won't try to talk you out of it.

Pick Nairi if any of these match: you want model and harness optionality (Claude Code, Codex, OpenCode, Cursor) instead of being locked to Claude; you want the same agent in Discord as well as Slack; you have compliance requirements that block cloud-only AI infra (data residency, IP allowlists, air gap); you want the option to self-host the agent runtime; or you care about owning the context layer separately from the model vendor. The shorter version: pick Nairi if you want to rent the intelligence and own the context.

Try Nairi in Slack or Discord

Install the Slack or Discord app, pick your channels, mention an agent. Two minutes, no infra to deploy. Pick the harness you want per agent. Keep the context.

Try Nairi free See Slack integration

Claude Tag and the context lock-in problem Nairi vs Claude Code for Slack Nairi vs Claude Managed Agents Nairi vs Cursor Cloud Agents Nairi vs Devin Nairi vs Lindy Nairi vs OpenClaw Nairi vs Viktor Slack integration Discord integration Pricing