fastpace Team used to be a feature list. Twenty-nine dollars a seat per month, bullet points underneath: org dashboard, SSO, SCIM, RBAC, compliance push, NHI export, redteam, model-eval, trust portal, webhooks, audit-log replication. Every item useful, every item defensible, none of it answering the question a VP of Engineering actually has when they land on the pricing page.
That question is not "what features do I get for twenty-nine dollars." It is "what problem are you solving that I currently can't solve myself."
We rebuilt Team to answer that question directly. This post is what changed, why, what it means for you if you are already on Team, and what is coming next.
The problem 5-to-50-dev orgs actually have
When an engineering organization grows past five developers, the AI tooling problem stops being individual and becomes organizational. The pattern is consistent enough that we can list it without hedging:
- Every developer's
.claude/directory drifts. The hooks one senior wrote in March are missing from the junior who joined in April. The agents one team uses are not the agents the other team uses. Nobody is wrong; everybody is slightly different. - Onboarding a new hire takes a senior developer's week. The senior walks the new hire through the codebase, the conventions, the agents, the slash commands, the local skills, the unwritten rules. The senior is not shipping during that week. The new hire is not shipping for two more.
- Nobody knows what AI is actually costing the company. The Anthropic invoice arrives, the OpenAI invoice arrives, the Gemini invoice arrives. They are line items, not insights. Which developer? Which repo? Which model? Which prompt patterns are wasted?
- The part-time security person — head of AppSec, fractional CISO, or the senior dev who cares about security on Fridays — has no visibility into what AI is doing in the developer environment. Their answer to "what's our AI dev governance" is a Confluence page that nobody reads and nothing enforces.
- The AI-generated code accumulates. It looks like any other commit. There is no audit chain, no signed manifest, no way to point at a regression and say "that was the model, here are the prompts, here is what changed."
Each of those problems has an obvious solution. The senior who wrote the hooks could publish them to a shared repo and write a sync script. The onboarding could become an internal wiki page. The cost dashboard could be a custom Grafana board against the Anthropic API. The security policy could be a YAML file in every repo. The audit chain could be a hook that signs commits.
They are all roughly a quarter of work for a senior engineer to build properly. Multiply by five capability areas and you have a year of one senior's time, spent not shipping product. Or you hire a platform engineer.
The three wrong answers
The 5-to-50-dev VP Eng has three options, and each is wrong in a different way.
Ignore it. Accept the drift. Hope the AI tooling sorts itself out. This is the most common answer because it is free. It is also the answer that compounds — drift accelerates, onboarding gets slower each hire, the surprise AI invoice gets bigger each quarter, and the security debt accrues until the day a customer's auditor asks the question that triggers the panic.
Hire a platform engineer. The honest number for a senior platform engineer in 2026: $250-350K base, $150-250K equity, ~30% overhead. Fully loaded, around $500K a year. Three of them, which is roughly what a real internal platform team for a 30-dev org looks like, is $1.5M a year. A 30-dev VPE who has had this conversation with their CFO knows exactly how it ends.
Build it yourself, between feature work. Pick one senior, give them 20% time, tell them to make the dotfiles consistent and the onboarding less painful. Six months later they have shipped one tool, the rest of the team has been blocked on review, and the senior is unhappy because they joined to build product, not internal tooling.
Each option fails in the same direction: the VPE knows the platform-team output is worth the investment, and cannot honestly justify the investment to get it.
The third option
fastpace Team is now the third option. The positioning sentence:
fastpace is the platform team your AI-coding surface needs — without the headcount.
Or, more concretely:
Everything a three-person platform team would build for AI-assisted development. Without the three people.
At $29 per editor-seat per month, billed annually, with the first five seats free for the first twelve months, a 30-developer organization spends roughly $10,400 a year on Team. Compared to one senior platform engineer at $500K a year fully loaded, the ratio is about 50 to 1. Compared to a real three- person platform team at $1.5M a year, the ratio is about 144 to 1.
A 50-to-1 cost ratio is the kind of number a VPE can take to a CFO without needing a deck. It is also the kind of number that raises a reasonable question: what is the catch.
The catch is that the framing only works if the product is true. A pitch that promises "platform team output" but ships "feature list" gets caught on minute three of the demo. So the rebrand is not just a positioning change. It is a forcing function on the product.
The four primitives that make it real
Real platform teams own a long list of things: standards, onboarding, knowledge, AI infrastructure, observability, quality, operations, security, workflow, culture. We mapped the catalog. There are roughly 40 features a platform team would build in their first two years for an AI-assisted- development surface. We are not shipping all 40 before the rebrand goes live; the catalog and sequencing are in our internal strategy doc.
Four are load-bearing. These are the table stakes — the four primitives that make "platform team in a box" true rather than aspirational. The rebrand does not go live until all four are in main and tagged.
| What a platform team would build | fastpace Team feature |
|---|---|
| Enforce one shared config baseline across every repo | fastpace sync |
| Onboard new hires in days, not weeks | onboarding agent (fastpace welcome) |
| Track AI cost per developer, per repo, per model | cost dashboard |
| Codify security policy your AppSec lead can defend | fastpace.policy.yaml |
fastpace sync is the most-requested missing primitive. One team
member updates a hook or a skill in the org-blessed .claude/;
every other developer's repo pulls the change on the next session. The
drift problem stops being a quarterly cleanup task and becomes a non-event.
The fastpace welcome onboarding agent takes a new hire from
"I just got my laptop" to "I shipped my first guided PR by Friday." It
reads the repo, knows the conventions, walks them through the agents and
slash commands, and adapts to their stated experience level. The two weeks
of senior hand-holding becomes three days, with a recorded transcript the
new hire can reread.
The cost dashboard is the procurement-justifying feature on its own. At 30 developers, organizational AI spend in 2026 is somewhere between $5K and $50K per month, and nobody knows where it goes. We surface the answer by developer, by repo, by model. Then we surface the waste: uncached calls, redundant prompts, oversized contexts. Most teams cut their AI spend 30-40% in the first two weeks just from seeing the breakdown.
fastpace.policy.yaml is the security beat that lets the
part-time security person defend the install. A single YAML file in the
repo declares which models are allowed, which MCP servers, the maximum
prompt size, allowed agent scopes, blocked paths. No DSL. No policy
engine to learn. Strong defaults. The CISO reads the file, the developer
reads the file, the auditor reads the file. They all agree on what is in
it because there is no interpretation layer in between.
Everything else Team includes — the org dashboard, SSO and SCIM and RBAC, compliance push to Drata and Vanta and Secureframe, NHI export, the redteam harness, model-eval, the trust portal, webhooks, audit-log replication — is now supporting evidence, not the headline. You are glad it is in the box. It is not the reason you bought the box.
What changes for current customers
Nothing gets taken away. Every feature that was in Team yesterday is in Team today. The four primitives are additions, not substitutions. The $29 price holds. The first-five-seats-free-for-twelve-months promotion holds. The score-pause guarantees hold.
Two things change.
First: fastpace sync, the onboarding agent, the cost dashboard,
and the policy file land in your install as part of the next release.
No new SKU, no upsell. If you bought Team for the dashboard and the SSO,
you now also get the primitives that make the dashboard and the SSO more
useful.
Second: we are introducing a new SKU above Team, called Platform, at $99
per editor-seat per month. Platform is for engineering organizations that
have already bought into Team and now want the differentiators a real
internal platform team would build in their first full year — cross-repo
refactor, DORA-AI metrics, team-aggregated learnings, fastpace ask
for cross-repo Q&A, an internal MCP server registry, incident replay,
a postmortem generator, drift detection, a prompt library, and custom
dashboards. We expect roughly 20-30% of Team customers will move to
Platform by the end of 2026. The other 70-80% will be perfectly served by
Team and we will not push them upward.
Enterprise is unchanged in feature surface — Platform-as-a-managed- deployment with the framework mappings included, MSA + DPA, named CSM, quarterly audit pack, plus the regulated-deployment add-ons. The Enterprise floor remains $25K per year or $99 per editor-seat, whichever is greater.
What is coming next
The four primitives are Wave 1 — the table stakes that have to be in main
before the rebrand goes live. Wave 2, shipping in two release waves after
the launch, is the differentiator catalog that opens the Platform SKU:
team-aggregated learnings, DORA-AI metrics, fastpace standup,
cross-repo codebase Q&A, the Slack and Linear integration, and the
cross-repo refactor agent.
Wave 3 is the heavier infrastructure that justifies the Platform price over a longer horizon: the model gateway, the MCP server registry as a curated marketplace, incident replay (uniquely ours because no other tool has the signed manifest chain), the postmortem generator, the drift detector that pairs with sync, and the prompt library.
Wave 4 is the long tail — roughly 20 additional features that are individually useful and collectively the second-year build of a real platform team. We will ship them opportunistically through 2027.
The thing we are not promising: that fastpace becomes the platform team for everything. The frame is specifically "platform team for the AI- coding surface." If your real platform team would also own Kubernetes, CI/CD pipelines, your dev-environment fleet, internal services — that is not us. We are not building Backstage. We are building the much narrower thing that no platform team existed for until AI-assisted development arrived.
The honest closing line
There is a version of this post that ends with a pitch and a CTA. The honest closing line is different:
If your engineering organization has more than five developers, you have the platform-team problem for your AI-coding surface, whether you have named it or not. The drift is real, the onboarding tax is real, the invisible AI invoice is real, the part-time security person's anxiety is real, the audit-chain gap is real.
The decision is not whether to solve the problem. The decision is whether to ignore it for another quarter, hire someone at $500K a year to solve it, build it yourself between feature work, or pay us roughly 2% of one of those salaries to deliver the same output. We think we have made that decision easy.
Either way, we would rather be honest about the choice than sell you a feature list.
Start with the Team tier (first five seats free for twelve months). If you have already outgrown the four primitives, the Platform tier is the upsell. If you are a bank, hospital, public-sector arm, or F500 division with regulated workloads, book a 30-minute Enterprise briefing.
