No. 01 The AI engineering OS

Your models ride in formation. When one tires, the next pulls through.

One workspace where your people and agents work together, on the Claude Code and Codex seats you already pay for. A person approves every change.

PullingClaude Code
Next upCodex
PaceHeld
Six models in formation. The task keeps moving.

Don't be left out when others are going at double speed.

Contact sales
Rides with
Claude CodeAnthropic APICodexOpenAI APIGemini APIAmazon BedrockVertex AIAzure AI FoundryOllama

01 Built like a paceline

The group rides faster
than anyone alone.

In a paceline, riders take turns at the front. When the leader tires, they peel off and the next rider pulls through, while the rest ride in the slipstream. That is how Paceline runs your work.

01

The pull

The right model at the front

Each step is done by the model it needs — small work on small models, hard work on strong ones — picked before an LLM token is spent.

02

The rotation

Limit reached? Next rider.

When a model hits its usage or rate limit, the next allowed one pulls through — on another vendor if it has to. Nobody waits for a reset.

03

The draft

Nobody starts from zero

The conversation, the approved plan, the staged edits and a handoff note ride along, so whoever takes the front is already in the slipstream.

04

The pacer

Someone holds the line

Each channel keeps its goal and roadmap, and the Pacer agent notices when the work drifts from it. Jira stays in sync.

02 The AI engineering OS

Meet the AI
engineering OS.

One place where your people, your agents and every model you already pay for work as a team — under one policy, with a person approving every change.

Who saves — and what

MoneyTimeProductivity

Stage 1

Startups & small teams

A small team with specialist agents as teammates, on the subscriptions you already have — no second AI bill to start.

MoneyProductivity

Stage 2

Scale-ups on several AI tools

Keep Copilot, Cursor and Claude Code. Put one policy and a bill per task over them, and carry on on another seat instead of buying overage.

MoneyTime

Stage 3

Enterprises

One governed layer over every vendor, with an audit trail your security team can export for review.

TimeProductivity

Stage 4

Regulated industries

Finance, health and the public sector: keep work on your own Bedrock, Vertex or Azure, or on local models.

TimeProductivity

Stage 5

Agencies & consultancies

A channel per client, each with its own budget and a bill per task, so AI spend lands on the right invoice.

MoneyTime

03 Race stats

Three numbers
your buyers will ask for.

Get the enterprise brief

CostFor the CFO

72%lower list price when each step runs on the model it needs — $0.695 became $0.193

  • A classifier decides, not a modelRoughly 1,500 tokens picks the model and thinking depth for each step, before an LLM token is spent.
  • Seats first, APIs as the fallbackRuns on the Claude Code and Codex subscriptions you already pay for.
  • A ceiling you can seeA token ceiling per task, a 30-day budget per channel, and a bill per task.

One measured task — build, design review, code review. Its actual token counts, priced at list twice: Opus for every step, and the Sonnet build and Haiku reviews Paceline chose. It measures choosing the model per step, not a run against another tool.

ProductivityFor the VP Engineering

0dead ends at "limit reached"

  • The task moves, not the developerWhen Claude hits a limit, the work continues on Codex, an API key or your company cloud.
  • Context travels with itThe conversation, the approved plan, the staged edits and a handoff note.
  • A team, not a toolPlanner, builder, designer and reviewer hand work on through approval gates.

Of the 23 AI coding products we researched, none document continuing a task on another vendor at a limit with its context.

SecurityFor the CISO

100%of writes approved by a person

  • Agents never write on their ownEvery edit is staged on a card a person approves. A file that changed since the agent read it is refused.
  • Secrets never reach a modelKeys and tokens become placeholders before any model call.
  • A log that shows tamperingAppend-only, each line carrying the hash of the one before.

Not yet certified: SOC 2 Type II is on the roadmap, with SAML and SCIM. Company sign-in with OpenID Connect works today.

04 The peloton

Your team and its
agents, in one room.

Channels where people and agents talk, plan and hand work to each other — and every change still waits for a person.

People and agents in one thread@mention any agent. It runs on that person's computer, with their models, and answers for everyone.
A shared library of agentsMake one once — role, instructions, skills — and use it in any channel. New ones wait for an admin.
Private channelsJust the people you name. Admins can see them, and the channel says so.
Files in the conversationScreenshots, text and PDFs on a message; agents read them as part of the thread.
Roles that mean somethingOwner, admin, member and guest. Guests read and comment.
A team server you hostCompany sign-in with OpenID Connect — Okta, Entra, Google Workspace. It holds messages, goals, agents and attached files, never your repositories or keys.
Goals and Jira beside the workEach channel keeps its goal and roadmap, and Jira stays in sync.

05 The course

Ask once.
Finish approved.

A person asks. Paceline decides the path. Nothing is written until someone approves it.

Channel

People and agents in one thread, with the goal and roadmap beside it.

Routing

Reads the request and sizes each step: model, thinking depth, which agents ride.

Agents

Planner → Builder → Designer → Reviewer, on the models your policy allows.

Approval gates

Plan, change, design and review each stop for a person.

Your repo

Written only when approved. Stale files are refused, never overwritten.

Before any model call: data policy · secret redaction · token ceilingAlways: an audit record of every model call, decision, approval and command

06 Time trial

Measured on a real task.
Then on yours.

Measured · 24 Sep 2026

Same tokens. The right model for each step.

Every step on the frontier model (Opus)$0.695
Each step on the model it needed (Sonnet build, Haiku review)$0.193

72% lower list price for the same work. One build → design review → code review task, run live in the desktop app; its token counts are what the models reported, priced at list both ways. This isolates the saving from choosing the model per step. It is not a head-to-head run against another tool — that is what the sandbox is for. Subscription seats are not billed per token; this shows relative cost.

Self-serve sandbox · in development

Run Paceline and Claude Code on your own repo.

Claude Codeon its own, the way your team uses it today
vs
Pacelinerouting, handoffs and gates on the same seats
  • The same repository and the same task, run side by side
  • Tokens, cost at list price and time to an approved change
  • Whether the tests pass, and the diffs each one produced
  • Both results reported — including a run Paceline loses
Book a benchmark on your repo

07 Standings

Where we stand —
and where we don't, yet.

Capabilities as each vendor documents them. We score ourselves by the same rule, and mark what is still on our roadmap.

Our reading of each vendor's official documentation, September 2026. "Partial" examples: Copilot audit-logs policy changes only; Claude Code redacts keys in telemetry only; Codex can pin but not enforce a model; Paceline has company sign-in with OpenID Connect today, with SCIM and SOC 2 still to come. Vendors change monthly — tell us if we are out of date.
CapabilityPacelineGitHub CopilotCursorClaude CodeOpenAI CodexDevin / Devin DesktopFactoryConductorClineWarp (Oz)
Models from several vendorsDocumentedDocumentedDocumentedNot documented or unverifiedPartial or limitedDocumentedDocumentedDocumentedDocumentedDocumented
Own cloud (Bedrock / Vertex / Azure) or local modelsDocumentedDocumentedDocumentedDocumentedDocumentedPartial or limitedDocumentedPartial or limitedDocumentedPartial or limited
Uses the team's existing Claude / ChatGPT subscriptionsDocumentedNot documented or unverifiedNot documented or unverifiedDocumentedDocumentedNot documented or unverifiedNot documented or unverifiedDocumentedNot documented or unverifiedPartial or limited
Keeps working on another vendor when a limit is hit, context keptDocumentedNot documented or unverifiedNot documented or unverifiedNot documented or unverifiedNot documented or unverifiedNot documented or unverifiedNot documented or unverifiedNot documented or unverifiedNot documented or unverifiedNot documented or unverified
Admin policy: which models and vendors may see codeDocumentedDocumentedDocumentedDocumentedPartial or limitedDocumentedDocumentedNot documented or unverifiedDocumentedNot documented or unverified
Secret redaction before the modelDocumentedNot documented or unverifiedNot documented or unverifiedPartial or limitedNot documented or unverifiedNot documented or unverifiedNot documented or unverifiedNot documented or unverifiedNot documented or unverifiedNot documented or unverified
Tamper-evident audit log of model callsDocumentedPartial or limitedDocumentedDocumentedDocumentedNot documented or unverifiedDocumentedNot documented or unverifiedPartial or limitedPartial or limited
Spend caps / token ceilingsDocumentedDocumentedDocumentedDocumentedPartial or limitedDocumentedPartial or limitedNot documented or unverifiedPartial or limitedPartial or limited
Human approval before every write and commandDocumentedDocumentedDocumentedDocumentedDocumentedDocumentedDocumentedPartial or limitedDocumentedNot documented or unverified
Team channels with planner / builder / reviewer agentsDocumentedPartial or limitedPartial or limitedPartial or limitedPartial or limitedDocumentedDocumentedPartial or limitedNot documented or unverifiedPartial or limited
SSO / SCIM and SOC 2Partial or limitedDocumentedDocumentedDocumentedDocumentedDocumentedDocumentedDocumentedPartial or limitedPartial or limited
Cloud / background agentsRoadmapDocumentedDocumentedDocumentedDocumentedDocumentedDocumentedPartial or limitedNot documented or unverifiedDocumented
● documented◐ partial or limited– not documented or unverifiedRoadmap planned for Paceline

Positioning · our assessment

Team orchestration (roles, gates, shared channels) → Team agents, one vendorGoverned, multi-vendor teams Individual toolsGateways (API-level)
  1. Paceline
  2. Devin
  3. Factory
  4. Warp Oz
  5. Conductor
  6. Claude Code
  7. GitHub Copilot
  8. OpenAI Codex
  9. Cursor
  10. Cline
  11. Goose
  12. LiteLLM
  13. OpenRouter

Governance across vendors (one policy, continuity, audit, redaction) →

Agents cannot write on their own

Every edit is staged on a change card a person approves.

Agents cannot run commands on their own

A command card shows exactly what will run, and runs only on approval.

Secrets never reach a model

Keys and tokens in files, chat or output become placeholders first.

One data policy per workspace

Company cloud, local, API keys and subscription CLIs, each allowed or blocked.

Tamper-evident audit trail

Each line carries the hash of the one before, so edits and deletions show.

Hardened desktop app

A sandboxed window, keys in the OS keychain, prompt-injection fencing.

What we don't have yet

We would rather you hear it from us.

The leading tools already carry certifications we don't. They are first on our roadmap:

  • SOC 2 Type II
  • SAML and SCIM (company sign-in with OpenID Connect works today)
  • Central policy from the team server
  • Audit streaming to your SIEM
  • Cloud and background agents

Until then, code and keys never leave a developer's machine except to the model vendors you have already approved, and every call is in the audit log.

08 The pilot

A 30-day
training block.

One team of five to fifteen, on your real backlog.

Week 1

Install and connect

Your existing Claude Code and Codex seats plus your company cloud. Set the data policy and the ceilings.

Weeks 2–3

Real backlog work

In channels, with a weekly readout: tasks, first-time approvals, tokens per task, cache share.

Week 4

Prove it

A security review of the exported audit log, and an ROI comparison against your current overage spend.

09 Pricing

Pay for the platform.
Keep your own models.

Priced per active engineer. Agents are unlimited, guests are free, and there's no markup on tokens: you bring your Claude Code and Codex seats, your API keys or your company cloud.

Team

One team, up to 25 engineers

Talk to usMonthly · per active engineer

  • Channels with Planner, Builder, Reviewer, Designer and Tester agents
  • Relay: when a model hits its limit, the task moves on with its context
  • The right model chosen for each step
  • Your models: Claude Code, Codex, API keys, Bedrock, Vertex, Azure, Ollama
  • Design Studio
  • Jira, GitHub, Slack, Trello and Microsoft Teams
  • A person approves every change before it's written
  • Usage by model, agent and channel; audit log kept 30 days
Book a pilot
Most teams

Business

25 to 200 engineers, several teams

Talk to usAnnual · 25-seat minimum

Everything in Team, plus

  • Company sign-in with Okta, Entra or Google Workspace
  • Admin policy: which models and vendors may see your code
  • Spend limits per team
  • Configurable secret redaction
  • Audit log export, kept for a year
  • Cost reporting by team, and the pilot ROI readout
  • Priority support and onboarding
Book a pilot →

Enterprise

200+ engineers, or regulated teams

Talk to usAnnual · custom contract

Everything in Business, plus

  • Self-hosted team server with deployment support and private networking
  • Security review pack: DPA, questionnaire and architecture
  • Data policy per project and custom log retention
  • Setup help for your company cloud
  • A named contact and an SLA
  • Volume pricing

On the roadmap: SAML and SCIM, SOC 2 Type II, cloud agents, and buying through the AWS, Azure and Google Cloud marketplaces.

Contact sales

Design partners: a free 30-day pilot for our first five teams, on your own backlog, with a week-4 readout of tokens per task and what you saved. In return we ask for a case study and a reference call.

Apply →

10 Questions

What buyers ask first.

We already have Copilot, Cursor or Claude Code.

Keep them. Paceline runs Claude Code and Codex as engines and adds what no single vendor can: continuity across vendors, one policy over all of them, and a team workflow with approval gates.

Is our code sent somewhere new?

No Paceline server sees your code. Agents run on each developer's machine, with the models your policy allows. The routing classifier receives only the request text, and an admin can turn it off.

Are you SOC 2 certified?

Not yet — it is first on the roadmap, alongside SAML and SCIM. Company sign-in with OpenID Connect (Okta, Entra, Google Workspace) and workspace roles already work on the team server. Meanwhile, code and keys never leave the developer's machine except to the model vendors you have approved, and every call is in the audit log.

What stops an agent from breaking things?

Nothing is written or run without a person's approval, a file that changed since the agent read it is refused rather than overwritten, and a runaway task stops at its token ceiling.

How is Paceline priced?

Per active engineer: anyone who sends a message or approves a change in the month. Agents are unlimited and guests are free. Team is billed monthly; Business and Enterprise are annual.

Do you resell tokens?

No. You bring your own models: your Claude Code and Codex seats, your API keys, or Bedrock, Vertex and Azure under your company's own contracts. We charge for the platform, with no markup on usage.

What does the pilot cost?

For our first five design partners, nothing: 30 days on your own backlog, with a week-4 readout. After that, a paid pilot is credited against the annual contract.

Can we buy through AWS, Azure or Google Cloud?

Not yet. Marketplace listings are on the roadmap so the purchase can count against your cloud commitment. Until then, we invoice directly.

11 Contact sales

Don't be left out when others are going at double speed.

Tell us about your team. We'll set up a 30-day pilot, run the benchmark on your own repository, and send you the enterprise brief.