One plan. Every model. One invoice.
Capuchin, Claude, Gemini, and ChatGPT in a single plan for every seat — with SSO, spend caps, audit-ready runs, and optional self-host. Your team gets every model without three separate subscriptions.
92.7%
of requests run on Capuchin
50+
agent tools implemented
~$0.06
avg per frontier request
100+
messages per session
Measured on production traffic.
Everything your admin needs. Nothing your developers don't want.
Governance, spend control, and compliance — layered on top of an IDE developers actually choose to use.
SSO and RBAC that fit your org.
One invoice. Every model. Hard caps.
Audit-ready by design, not by retrofit.
Self-host everything — or nothing.
Teams that can't compromise on control — or speed.
Whether you're a 10-person startup or a 500-engineer enterprise, MonkeysCode adapts to your governance needs without slowing anyone down.
Regulated enterprises
Signed audit trails, air-gapped deployment, and no-training guarantees satisfy compliance teams before they have to ask.
ComplianceFast-moving startups
One plan covers every model — no juggling three subscriptions. Shared allowance pooling means the team that ships more gets more.
VelocityOpen-source maintainers
Open core, Open VSX, no lock-in. Run it fully local or BYOK — your contributors keep their workflow even after they leave.
FreedomPolyglot engineering orgs
First-class PHP, Python, Go, Rust, and TypeScript. Monorepo-aware graph so agents reason across services, not just nearby files.
PolyglotWhat your team is paying now, versus one invoice.
A twenty-person engineering team, currently doing what most teams do.
| Today | With MonkeysCode | |
|---|---|---|
| AI editor subscription | 20 seats, billed separately | Included |
| Claude API | Separate account, metered | Included in allowance |
| OpenAI API | Separate account, metered | Included in allowance |
| Gemini API | Separate account, metered | Included in allowance |
| Invoices to reconcile | Four | One |
| Spend predictability | Metered, variable | Hard cap |
| Admin visibility | Per-provider dashboards | One dashboard |
Everyday work runs on unlimited Capuchin. Frontier tokens are reserved for the problems that deserve them — ~$0.06 average per frontier request, measured, because the on-device context engine sends relevant code slices rather than whole files.
One subscription per seat. Every model included.
Unlimited Capuchin on every plan. Claude, Gemini, and ChatGPT draw from a shared team allowance — no per-model surcharges, no surprise invoices.
Team
For small teams that want every model in one plan. Up to 50 seats.
Start free- IDE, Code Agent and CLI included
- Unlimited Capuchin (Flash + Reason) per seat
- $30/mo frontier allowance (Claude/Gemini/ChatGPT)
- Admin dashboard with seat management
- Centralized billing — one invoice
- Hard spend cap, per-seat budgets
- Shared encrypted index (opt-in)
- Signed, replayable agent runs
- All 50+ agent tools
Team Pro+
For growing orgs that need more headroom and governance. Up to 100 seats.
Start free- Everything in Team, plus:
- IDE, Code Agent and CLI included
- $50/mo frontier allowance
- Priority inference queue
- SSO (SAML / OIDC)
- Role-based access control (RBAC)
- Team-wide model routing policies
- Audit log export (SIEM-ready)
Team Ultra
For heavy agentic workloads at scale. Up to 200 seats.
Contact sales- Everything in Team Pro+, plus:
- IDE, Code Agent and CLI included
- $80/mo frontier allowance
- Maximum parallel and background agents
- Granular spend caps per team / project
- Priority support
Enterprise
For regulated industries and air-gapped environments. Unlimited seats.
Contact sales- Everything in Team Ultra, plus:
- IDE, Code Agent and CLI included
- Air-gapped / fully on-prem deployment
- Self-hosted model endpoints (vLLM / Ollama)
- Custom data residency agreements
- Dedicated success engineer
- SCIM provisioning
- GDPR-ready architecture with data residency controls
- Contractual SLA available
Every plan includes unlimited Capuchin, signed replayable runs, and zero telemetry by default.
BYOK supported on all tiers — bring your own Claude, Gemini, or ChatGPT keys.
Your code stays yours. Always.
Zero
IDE telemetry by default
No tracking, no training
The IDE collects zero telemetry by default. Opt-in anonymous crash reporting only. Our marketing website uses standard analytics.
100%
Local or self-hosted
Air-gapped deployment
Local index, local models, internal registry. Zero outbound calls. Runs behind your firewall with no internet egress.
Every
Agent run signed
Audit-ready trails
Tamper-evident event logs for plan, tools, and diffs. Replayable byte-for-byte for any reviewer or compliance officer.
The same engine, everywhere your team works.
IDE, Code Agent, and CLI — all included in your plan. Four execution environments per surface.
Code Agent
Launch, watch and orchestrate parallel agents across projects without opening the editor. Each agent isolated, each producing a signed diff your team reviews before merge. Same 50+ tools, four execution environments.
Learn moreMonkeysCode CLI
Agentic coding from the command line and CI. Headless, structured output, real exit codes. Runs Capuchin, Claude, Gemini, ChatGPT or your own model. The same permission gating and signed runs as the IDE.
Four backends, governed by your admin.
Each backend declares its capabilities honestly. Admins can restrict which environments are available per team or project — for example, allowing only Cloud and SSH in production workspaces while permitting Local for personal development.
The Code Agent is available now. CLI arrives on existing plans at no additional cost.
Questions teams ask.
How does the shared model allowance work for teams?⌄
Every seat gets unlimited Capuchin. Claude, Gemini, and ChatGPT draw from a shared team-wide monthly pool. When the pool is exhausted, Capuchin continues running uninterrupted — only frontier model calls pause until the next cycle. Admins can raise the pool or switch to BYOK at any time.
How does Capuchin compare to the frontier models on quality?⌄
Capuchin is specialised for software engineering rather than general conversation, which is why it handles 92.7% of production requests in real teams while frontier models are reserved for the hardest 7%. We benchmark it on our own production endpoint and publish the methodology. Every plan includes Claude, Gemini and ChatGPT for the work that genuinely needs them, so the choice is never forced.
Can we enforce which models our developers use?⌄
Yes. Team Pro+ and higher plans include team-wide model routing policies. Admins can allowlist or blocklist specific models, enforce Capuchin-first routing, or require BYOK for certain project tags. Policies apply transparently — developers see the active routing in their model picker.
What does "signed, replayable runs" mean for compliance?⌄
Every agent run produces a tamper-evident event log: the plan, every tool invocation, every diff, and the final result — all cryptographically signed. A compliance officer or auditor can replay the run byte-for-byte to verify what happened. This is built into every plan, including Free.
Can MonkeysCode run fully air-gapped?⌄
Yes. In air-gapped mode, the IDE uses a local on-device index, local models (Ollama, llama.cpp, vLLM), and an internal extension registry. Zero outbound network calls. Enterprise customers can deploy this behind a firewall with no internet egress whatsoever.
Do you train on our code or prompts?⌄
No. Your codebase, prompts, and agent outputs are never used for model training — not for Capuchin, not for any provider. The IDE collects zero telemetry by default; opt-in anonymous crash reporting only. Our marketing website uses standard analytics. This is a product-level guarantee, not a config toggle.
How does pricing scale as we add seats?⌄
Team is $30/seat/month, Team Pro+ is $50/seat/month, Team Ultra is $80/seat/month, and Enterprise is $200/seat/month — all with unlimited Capuchin per seat and a shared frontier allowance. There is no per-model surcharge on any plan.
We already have API keys for Claude / Gemini / ChatGPT. Can we use those?⌄
Absolutely. BYOK is supported on every plan, including Free. Teams can mix managed allowances (included in the plan) with BYOK keys (billed to your provider) — admins can even set policies for which approach each team uses.
What's the onboarding process for a team of 50+?⌄
After SSO is configured, new developers are auto-provisioned via SCIM. The admin dashboard lets you assign seats, set budgets, and configure model routing before anyone writes their first prompt. Our team offers guided onboarding sessions for Team Pro+ and Enterprise customers.
Give your team every model.
One plan. One invoice. Zero lock-in.
Start free with Capuchin — no card. When your team is ready, upgrade to a plan that includes Claude, Gemini, and ChatGPT for every seat, with the governance your org needs.