Does LaunchDarkly include audit logs on Foundation?

No. Audit logging, approvals, SCIM, and custom roles are Enterprise, custom pricing, a sales conversation. That is the compare table as of mid-2026. Do not confuse this with SSO: LaunchDarkly's own pricing page has listed SSO on Foundation in some places, so treat SSO as messy and do not build an argument on it.

Why do audit logs matter for feature flags?

A flag flip can take checkout down or bring it back. After an incident you need to know who changed what, when, and from where, including which agent key if a coding agent did it. Without an audit log, a kill switch is a rumor.

Does Betterflag require Enterprise for audit logs?

No. Audit log, RBAC, and SSO are on the Scale tier at $99.99/mo, self-serve, published price. No sales call. Approvals for sensitive agent actions (production kill switches) live there too.

Is LaunchDarkly still worth it without the audit log?

For a large enterprise that will buy Enterprise anyway, yes. For a startup on Foundation, you have a kill switch you cannot forensically explain. That is a strange place to save money. Either buy the tier that has the log, or use a vendor that puts the log on a public price.

LaunchDarkly Audit Logs Are Enterprise-Only. That Is Backwards.

A kill switch that nobody can explain after the fact is a rumor. Checkout went down. Someone turned a flag off. Or on. Or an agent did. You will be asked who, and "we think it was in the dashboard" is not an answer a serious team gets to give twice.

LaunchDarkly knows this. That is why audit logging, approvals, SCIM, and custom roles are Enterprise: custom quote, sales call, not the Foundation card. The safety mechanism is the upsell.

I am building Betterflag. This is the second live hook after you are billed for your architecture. Verify against their pricing page. The side-by-side is Betterflag vs LaunchDarkly.

What is actually gated (and what I will not claim)

As of mid-2026, the compare table puts these on Enterprise:

Audit log. Who changed the flag.
Approvals. A second pair of eyes before prod moves.
SCIM. Provisioning that is not a shared password spreadsheet.
Custom roles. Something narrower than "everyone is admin."

I am not going to claim SSO is Enterprise-only. LaunchDarkly's own page has disagreed with itself: Foundation cards have listed single sign-on while the compare table showed it as missing. Screenshot the Foundation card and a "SSO is Enterprise" post is dead. Build on the unambiguous rows. Audit log, approvals, SCIM, custom roles.

Unlimited seats on Foundation does not help you if the log of what those seats did is a different SKU.

Why this is backwards for flags

Flags are not analytics. A wrong flag is an incident. The minimum governance for that is:

Every change is recorded (human or agent).
Production kill switches can require a second human.
You can revoke one actor without rotating the company.

That is not "enterprise." That is Tuesday. Putting it behind custom pricing means startups on Foundation run kill switches as folklore, and enterprises pay a tax to get a paper trail that should have been in the boolean product.

Approvals are the same story. An agent that can kill_flag in production without a human tap is a scary default. An agent that cannot be audited is worse. Guardrails without a log are theater.

What "good" looks like on a public price

Betterflag Scale is $99.99/mo, self-serve:

Audit log (who, or which agent key, flipped what, when)
RBAC
SSO
Approval guardrails for sensitive agent actions

No "contact sales." No custom quote to find out if the log exists. Starter and Launch already include unlimited seats and agent keys; Scale is the governance tier, not a personality change.

If $99.99 is too much and you truly do not need a log, you can live on Launch ($24.99) and accept that forensics will be thin. I would rather you know that than discover it at 2am. I would rather you not discover it on a vendor that hid the log behind a sales org.

Agents make this less optional, not more

Coding agents now create flags and stage rollouts. If they authenticate as you, the audit line says you flipped checkout at 2am when you were asleep. Agent-scoped keys plus an agent-attributed log are how this stays sane.

LaunchDarkly has an MCP server, including on free. MCP existence is table stakes in 2026. The question is whether the write path is first-class (full API parity, scoped keys, approvals, log) or a dashboard with an API in the sidebar. I care about the second list. You should too if Claude Code is already opening PRs you only skim.

If you are staying on LaunchDarkly

Buy Enterprise if you need that governance, and go in with eyes open on connections and MAUs. Do not sit on Foundation and tell yourself you have a kill switch you can defend. You have a toggle.

If you are leaving, you want three things in the replacement: the four primitives (flags 101), one meter you can forecast, and a log that is not a quest. That is the product.

Betterflag publishes the log on Scale. Alpha waitlist members lock in 50% off for life. Join if you have ever had to guess who flipped the flag.