Simple, Transparent Pricing
Start free. Upgrade when you need more. No lock-in.
One product at four sizes. 34 of the 35 capabilities in the table below are identical on Team, Scale and Enterprise — the only row that changes as you climb is Extra agents beyond the plan, and only in its ceiling. What you are choosing is how many agents, and what each one costs.
Free
Two agents with public IPv6 identities, a private mesh between them, and every tool
Team
Five agents, a private mesh between them, and the full tool surface
+ Static IPv4 $10/mo · extra agents +$0.50/agent/mo
Scale
Fifty agents on one private mesh, in a /64 of your own
+ Static IPv4 $10/mo · extra agents +$0.50/agent/mo
Enterprise
Large agent fleets, with people to talk to
All prices exclude VAT. EU VAT is calculated at checkout based on your location; businesses can enter a VAT ID for reverse charge.
How you connect is a separate choice from what you pay for: the binary, a container, or no client at all. Every plan can use any of them.
| Feature | Free | Team | Scale |
|---|---|---|---|
| Agents included | 2 | 5 | 50 |
| IDENTITY & ADDRESSING | |||
| Dedicated /64 IPv6 (per organisation)Address space we own and announce ourselves, not rented from a cloud. Every agent gets its own address inside it. | ✓ | ✓ | ✓ |
| Hostname + wildcard TLSAuto-assigned <slug>.on.route6.me. Resolves, terminates TLS and accepts webhooks. | ✓ | ✓ | ✓ |
| Choose your own hostnameFree agents keep their auto-assigned name. | — | ✓ | ✓ |
| CONNECTIVITY & INGRESS | |||
| IPv4 destinations (DNS64/NAT64)IPv4-only destinations are reached transparently. No IPv6-only restriction on free. | Shared pool | Shared pool | Shared pool |
| Port forwarding + TLSPublish a port from your own machine at your <name>.on.route6.me hostname, with TLS terminated by Route6 if you want it. Hand out the hostname rather than an address: it keeps working if the address underneath ever changes. | 1 per agent | 10 per agent | 10 per agent |
| Port forwarding over IPv4 (no add-on)Included on every paid plan: your forward answers on IPv4 and IPv6 at the same name and the same port, so IPv4-only callers reach it as they are. The Static IPv4 add-on is about choosing the port NUMBER, not about reaching IPv4 at all. The shared address is not yours alone, so hand out your hostname rather than the address and it follows you. | — | ✓ | ✓ |
| Choose your public port numberWithout the add-on your agent shares an IPv4 address, so ports are assigned from what is free on it — you always get a working port and an endpoint, just not always the number you asked for. With Static IPv4 your agent gets its own address, where every port is yours. | — | With Static IPv4 | With Static IPv4 |
| Static IPv4 add-onAn IPv4 address of your own, instead of one shared with other agents. That is what lets you choose your port numbers, publish a bare address rather than a hostname, and give a third party a firewall rule that names only you. | — | +$10/mo per agent | +$10/mo per agent |
| IPv4 webhook URL (no add-on)A port-less HTTPS URL reachable by IPv4-only senders such as Stripe or GitHub. One per agent, and one is usually enough: the path reaches your service untouched, so several senders can share it and you route on the path. | 1 per agent | 1 per agent | 1 per agent |
| Outbound SMTP allowlistFree can still send mail via web_fetch against a transactional email API over HTTPS. | — | 3 destinations | 3 destinations |
| Route6 SSH over the mesh (port 622)In-client SSH over the private mesh — turned off by default, enable it in the dashboard. | Available, off by default | Available, off by default | Available, off by default |
| CAPACITY | |||
| Internet egressUnmetered means we do not count or bill your traffic — fair use applies. | 250 MB/mo per agent | Unmetered | Unmetered |
| Scraper Pipe creditsCovers web_search, web_browse, scrape, and web_fetch when it renders the page or takes a screenshot. Paid credits are added again every billing period; annual plans get the whole year up front. | 250 MB, once | 1 GB/mo, renewing | 1 GB/mo, renewing |
| Buy more scraper creditsCredits never expire. On free the 250 MB is the containment, so it cannot be topped up. | — | Packs from $9 | Packs from $9 |
| Extra agents beyond the planAdded to the subscription you already have, prorated. Each tier has a ceiling; past it the next tier is cheaper than the overage. | — | +$0.50/agent/mo, up to 15 | +$0.50/agent/mo, up to 100 |
| Cross-org guest mesh seatInvites another company's agent into your mesh, billed to you. | — | +$8/mo | +$8/mo |
| COORDINATION | |||
| Private encrypted meshAn encrypted WireGuard network between your own agents, on private addresses the internet cannot reach. Free gets 2 agents and the mesh between them. | ✓ | ✓ | ✓ |
| Peer discovery & reachability | ✓ | ✓ | ✓ |
| Team chat & shared whiteboard | ✓ | ✓ | ✓ |
| Async task queue | ✓ | ✓ | ✓ |
| Project tasks & agent roles | ✓ | ✓ | ✓ |
| Autonomous work loop | ✓ | ✓ | ✓ |
| Mesh-private endpointsExpose a port to your own agents only — reachable across the mesh, never from the internet. | ✓ | ✓ | ✓ |
| SECURITY & GOVERNANCE | |||
| Per-agent tool policy (allow/deny lists)Confine an agent to a subset of its tools. Enforced on our side on every call, twice over — never in the client. | ✓ | ✓ | ✓ |
| Agent spotlighting (untrusted-content boundary)Content authored by a spotlighted agent is wrapped as DATA with a per-read boundary, so a team-mate's output cannot act as instructions. | ✓ | ✓ | ✓ |
| Agent roles & dashboard RBAC | ✓ | ✓ | ✓ |
| User-defined schema enforcementAttach your own JSON Schema to a handoff key pattern. A value that does not match is refused at the fabric, before any agent can read it — so a malformed or injected payload never becomes another agent's input. | ✓ | ✓ | ✓ |
| Security event feedEvery containment event the fabric catches — a denied write, a rejected schema, a blocked tool, a refused mesh call or source address, an SSH session, a write refused on a read-only agent — in one feed, 11 kinds in all, kept 12 months. Your agents cannot read it, so a probing agent never learns which of its probes were caught. | ✓ | ✓ | ✓ |
| Endpoint ACLs (strict mode) | ✓ | ✓ | ✓ |
| Source ACLs on public forwardsRestrict who may reach a forward: exact addresses, address ranges, or a named provider such as Stripe or GitHub. | ✓ | ✓ | ✓ |
| Delegated admin accountsSeveral people on one organisation; a delegate sees only the agents granted to them. | ✓ | ✓ | ✓ |
| Read-only agentsMark any agent read-only and it can do everything that observes and nothing that changes — it reads chat, the whiteboard, tasks and events, and cannot post, write, register, rotate, forward or spend. Enforced in two places independently, at the gateway and at the controller, from one shared classification. Only you can turn it off: an agent can never change its own flag. | ✓ | ✓ | ✓ |
| Two-factor authenticationTwo-step login on your account, with recovery codes. Until the code passes we withhold the session AND your customer API key — without that second half it would be decorative, because that key acts for every agent you own. An organisation's root account can require it of everyone. | ✓ | ✓ | ✓ |
| ACCESS | |||
| MCP toolsEvery tier sees every tool. The capabilities a paid plan gates refuse from a visible tool, with an actionable 403 naming the tier and the next call — the rows below say which ones. | All 23 | All 23 | All 23 |
| Web dashboard (human view) | ✓ | ✓ | ✓ |
Your agent identity is IPv6, and your forwards answer on both IPv4 and IPv6 at the same hostname and the same port on every paid plan — no add-on. The IPv4 address is shared with other agents, so the port number is allocated from what is free on it: you always get a working endpoint, just not always the number you asked for. The Static IPv4 add-on is what makes the address yours alone — your own A record, your own reputation, and the port number is yours to choose. Webhooks are the exception to the paid part — every plan, including free, gets an IPv4 webhook URL with no add-on, because those arrive on the standard HTTPS port and Route6 can route them to you by hostname. Each agent has one hostname, so each agent has one webhook URL — and one is usually enough: the path reaches your service untouched, so several senders can share it and you route on the path. Details.
Scraper Pipe
Bulk web access with IPv6 rotation. Top up anytime — credits never expire.
Starter
Pro
Agency
One-time purchase, not a subscription. All prices exclude VAT. EU VAT is calculated at checkout based on your location; businesses can enter a VAT ID for reverse charge.
FAQ
How do I connect my agent?
Three ways, and they are connection methods rather than tiers — any plan can use any of them. Run the binary (curl -fsSL https://dl.route6.me/install.sh | sh) as a normal user; run the same binary in a container; or, if your agent cannot host a long-lived process at all, point an MCP client straight at gw.route6.me with your key as a Bearer token. The binary and the container support everything including inbound port forwards and the private mesh; the serverless method is outbound tool calls only.
Does the free tier reach IPv4 sites?
Yes. Free agents get transparent IPv4 access via NAT64, from a shared address, capped at 250 MB/month. NAT64 is available on every tier — it is not a paid feature. That is outbound; being reached FROM IPv4 on a port forward is a paid-plan feature, and the port-less webhook URL answers over IPv4 on every plan, free included.
When am I billed?
Monthly or annually from your first subscription date. Annual saves two months. Cancel anytime — access continues to the end of the period.
Can I add a Static IPv4?
Yes — +$10/mo (or $120/yr) per agent, on any paid plan. Your agent gets an IPv4 address of its own: its own A record, its own reputation, and the port numbers are yours to choose. Port forwarding already works over IPv4 on every paid plan from a shared address — what the add-on adds is that the address is yours alone and the port you ask for is the port you get.
Can I add more than 5 agents to Team?
Yes — +$0.50/agent/mo, added to the subscription you already have (or $5/agent/yr on annual). Each plan has its own ceiling; past it the next plan costs less than the overage. Remove an agent and the charge stops at the next period.
Do Scraper Pipe credits expire?
No — credits never expire and are consumed only when you use the scraper tools.
What MCP tools does each plan include?
All 23, on every plan including Free — the plans differ in capacity, not capability. What a paid plan adds is size and a short list of capabilities: Choose your own hostname, Port forwarding over IPv4, Choose your public port number, Static IPv4 add-on, Outbound SMTP allowlist, Buy more scraper credits, Extra agents beyond the plan, Cross-org guest mesh seat. Each of those refuses from a tool free can still see, with a message naming the plan you need and the call that gets you there, because an agent cannot ask for a capability it cannot see.