Add token source API for fetching LiveKit tokens by alan-george-lk · Pull Request #177 · livekit/client-sdk-cpp

alan-george-lk · 2026-06-18T00:10:20Z

Summary

Adds new public API token_source.h, allowing users to now mint tokens using the following paths:
- Literal: user-provided token (basically the legacy livekit-cli minted path)
- Sandbox: enable a sandbox token server via LiveKit cloud, and allow users insecurely mint tokens (development only)
- Endpoint: production-level server support
Additional public API changes include:
- room.h -> new ways to connect with tokens
- room_delegate.h -> new callback with the new event
- room_event_types.h -> new information-only event type when tokens are refreshed by the server (see below)

Server Token Refresh

LiveKit server token refresh existed before this feature, but this branch now adds support for the FFI event onTokenRefreshed.

New Dependencies

libcurl4-openssl - used for HTTP token endpoint support
nlohmann/json - JSON serialization/deserialization support for all configurable token source types

Testing

Validated through a combination of unit, integration, and standalone binary tester program against a LiveKit Cloud agent with a live token server.

Token Server Action

Integration tests now test against a basic token server implementation run in GitHub actions live (skipped locally).

Sandbox Token

Verified locally by connecting to a LiveKit cloud instance:

auto sandbox_token = livekit::SandboxTokenSource::fromSandboxId("<omitted but valid>");

And results printed below:

./build-release/bin/token_source_tester
...
Server URL: <omitted but valid>
Participant Token: <omitted but valid>
Participant Name: efficient-sensor
Connected to room
Local participant: cpp-test-a

xianshijing-lk

might need help to understand the TokenSource a bit more

…usage)

Manual ad-hoc helper for exercising token sources locally; the sandbox ID must be supplied via LIVEKIT_SANDBOX_ID rather than hard-coded in source. Co-authored-by: Cursor <cursoragent@cursor.com>

MaxHeimbrock · 2026-06-23T09:00:35Z

I am using the tokensource sandbox and it is working fine for me 👍

…/token_source_api

xianshijing-lk

lgtm with some comments, please address them.

Great work!!!

xianshijing-lk · 2026-06-25T13:44:59Z

@@ -0,0 +1,109 @@
+# Token lifecycle
+
+Succinct reference for how join credentials and in-session refresh interact in


@1egoman , could you please help review this token-lifecycle.md ?

I actually generated this for him while I had some active LLM context around this topic, I removed it as I don't want this SDK to be an authoritative documentation source compared to docs.livekit.io.

I do think we should have some better docs around token source and lifetimes, while going through this process it felt a little disjointed and agents/frontend-focused and not just a general document for token source terms, functions, lifecycles, etc for client SDKs.

xianshijing-lk · 2026-06-25T13:47:40Z

 #include "livekit/room_event_types.h"
 #include "livekit/stats.h"
 #include "livekit/subscription_thread_dispatcher.h"
+#include "livekit/token_source.h"


nit, would you consider forward declare those types, and move the #include "livekit/token_source.h" to the cpp ?

One benefit for it is that it could speed up the build.

xianshijing-lk · 2026-06-25T13:52:53Z

+/// @brief Base interface for token sources that provide full credentials directly.
+class LIVEKIT_API TokenSourceFixed {
+public:
+  virtual ~TokenSourceFixed();


nit, just do "virtual ~TokenSourceFixed() = default" here and remove the cpp line

The same for virtual ~TokenSourceConfigurable() = default;

xianshijing-lk · 2026-06-25T14:04:53Z

+  return connect(details.value().server_url, details.value().participant_token, options);
+}
+
+bool Room::connect(TokenSourceConfigurable& token_source, const TokenRequestOptions& request_options,


consider a helper function like

namespace { template <typename FetchFn> bool connectWithTokenSource(Room& room, const RoomOptions& options, FetchFn&& fetch) { Result<TokenSourceResponse, TokenSourceError> details = Result<TokenSourceResponse, TokenSourceError>::failure(TokenSourceError{"token source not invoked"}); try { details = fetch().get(); } catch (const std::exception& e) { LK_LOG_ERROR("Room::connect failed: token source threw: {}", e.what()); return false; } catch (...) { LK_LOG_ERROR("Room::connect failed: token source threw unknown exception"); return false; } if (!details) { LK_LOG_ERROR("Room::connect failed: token source error: {}", details.error().message); return false; } const auto& value = details.value(); return room.connect(value.server_url, value.participant_token, options); } } // namespace

Then these two function can become:

bool Room::connect(TokenSourceFixed& token_source, const RoomOptions& options) { return connectWithTokenSource(*this, options, [&] { return token_source.fetch(); }); } bool Room::connect(TokenSourceConfigurable& token_source, const TokenRequestOptions& request_options, const RoomOptions& options) { return connectWithTokenSource(*this, options, [&] { return token_source.fetch(request_options, false); }); }

FWIW, we opted not to do this in other sdks, and determined this was the session api's job. Relevant pull request: livekit/client-sdk-js#1677

xianshijing-lk · 2026-06-25T14:23:31Z

+    : provider_(std::move(provider)) {}
+
+std::future<Result<TokenSourceResponse, TokenSourceError>> CustomTokenSource::fetch(const TokenRequestOptions& options,
+                                                                                    bool /*force_refresh*/) {


any idea why CustomTokenSource will skip the force_refresh ?

This comment actually led to an API design change, see here: e39e55e

Having such an argument for a custom source would require passing that argument along to the custom implementation, which felt weird as the custom token source was just a user function that could decide anything it needed to. I have instead dropped force_refresh entirely, and opted for caching_token_source.invalidate() which matches how Swift/Android do it.

In short, this original design around forcing refresh was modeled after JS with the force_refresh. But JS actually goes for an inheritance-based approach to cacheing token sources, vs. here which is using a decorator (Swift/Android do it this way), so forcing refresh was consistent across all sub-classes and didn't have an awkward drop like this did.

It was basically a divergence in both approaches (force + decorator vs. force + inheritance), now it's more aligned to Swift/Android.

xianshijing-lk · 2026-06-25T14:25:20Z

+      return TokenSourceResult::success(*cached_details_);
+    }
+
+    auto result = inner_->fetch(*options_snapshot, force_refresh).get();


any reason why we want to run the inner_->fetch() under the lock ?

Will it be safer to do ?

{ std::scoped_lock lock(mutex_); if (!force_refresh && cached_details_ && cached_options_ && tokenRequestOptionsEqual(*cached_options_, *options_snapshot) && isParticipantTokenValid(cached_details_->participant_token)) { return TokenSourceResult::success(*cached_details_); } } auto result = inner_->fetch(*options_snapshot, force_refresh).get();

Holding the mutex over the fetch is how JS and Swift do it, I'm open to your change but it would be a divergence

xianshijing-lk · 2026-06-25T14:28:03Z

+  auto source = std::unique_ptr<SandboxTokenSource>(new SandboxTokenSource(sandbox_id, options, base_url));
+  auto resolved = resolveSandboxEndpoint(sandbox_id, std::move(options), base_url);
+  source->endpoint_ =
+      EndpointTokenSourceTestAccess::create(std::move(resolved.url), std::move(resolved.options), std::move(transport));


isn't the endpoint_ created in line 210 already ? any reason why it needs to be re-created ?

Great catch, this slipped through. Now have a private constructor that accepts an already made endpoint so both the public and test access versions can avoid recreation

xianshijing-lk · 2026-06-25T14:33:20Z

+  }
+
+  const std::wstring host(components.lpszHostName, components.dwHostNameLength);
+  const std::wstring path(components.lpszUrlPath, components.dwUrlPathLength);


question, does path need to include lpszExtraInfo ? like when lpszExtraInfoLength > 0 ?

…/token_source_api

1egoman

Generally looks good to me, and I think matches the interfaces on other sdks pretty well!

1egoman · 2026-06-25T18:35:41Z

+  /// LiveKit Cloud deployment to target for agent dispatch.
+  ///
+  /// Optional. When omitted or empty, the production deployment is used.
+  /// Only relevant when dispatching a named agent on LiveKit Cloud.
+  std::optional<std::string> agent_deployment;


nitpick: in other sdks (for example, web) this was just called deployment. I like this agentDeployment name better though and had brought it up during the review process when it was added recently but I think it got missed. It might be worth changing this for consistency, idk.

Android and Swift (which this largely models) both use agentDeployment. So I think it's a 50/50 consistency roll of the dice 🤔

Ugh I hadn't realized it was so fragmented 😧

1egoman · 2026-06-25T18:45:10Z

+bool tokenRequestOptionsEqual(const TokenRequestOptions& a, const TokenRequestOptions& b) {
+  return a.room_name == b.room_name && a.participant_name == b.participant_name &&
+         a.participant_identity == b.participant_identity && a.participant_metadata == b.participant_metadata &&
+         a.participant_attributes == b.participant_attributes && a.agent_name == b.agent_name &&
+         a.agent_metadata == b.agent_metadata && a.agent_deployment == b.agent_deployment;
+}


nitpick: I'm not sure if there's a way to handle this in c++, but ideally if possible it would be great if the compiler could somehow throw an error if a field in a or b were missed in this check.

Also, do you have to worry about deep equality here within maps like participant_attributes? Or is == on the fields directly good enough?

For the second part, == is an "operator overload", which is a C++-ism that basically allows you to define actual functions for operators like +, -, *, ==, etc., so for maps it is in fact a deep comparison: https://en.cppreference.com/cpp/container/map/operator_cmp (good callout though!)

For the first one, there are some options, I am looking into that

@1egoman For the first part, it's a good callout, but until C++20 which now has default bool operators (compiler-generated field quality checks like this does manually), what you see here is basically the way to do it unfortunately. There's some C++17 (what we're on) features to slightly improve the robustness, but it adds a lot of code for little benefit.

My vote would be to leave this as-is as it's unlikely to change often

1egoman · 2026-06-25T18:51:14Z

+/// @brief Decorator that adds JWT-aware caching to another configurable token source.
+///
+/// Wrap @ref CustomTokenSource, @ref EndpointTokenSource, or
+/// @ref SandboxTokenSource to reduce token fetch calls while still refreshing
+/// when tokens expire or when @p force_refresh is requested.
+class LIVEKIT_API CachingTokenSource final : public TokenSourceConfigurable {


thought: You probably have seen it / maybe even patterned it off of this, but your implementation here is fairly close to the swift version: https://github.com/livekit/client-sdk-swift/blob/5d13e9ca7b366a8b47bd08c95a34cb0068149287/Sources/LiveKit/Token/CachingTokenSource.swift#L23.

Also just want to confirm - looking at the cpp I think this is just memoization with the last value correct? There's no other way to configure this right? I think that fine.

Yup and yup to your two questions!

xianshijing-lk reviewed Jun 18, 2026

View reviewed changes

Comment thread include/livekit/room.h Outdated

alan-george-lk force-pushed the feature/token_source_api branch 2 times, most recently from 9e60f34 to 1929f4a Compare June 22, 2026 16:44

github-advanced-security AI found potential problems Jun 22, 2026

View reviewed changes

Comment thread .github/workflows/make-release.yml Fixed

Comment thread .github/workflows/make-release.yml Fixed

alan-george-lk and others added 12 commits June 22, 2026 13:51

Initial token source API

85504d0

Add integration test

df603b6

Actual TokenSource logic

4bc2b61

Cleanup for clang/CI build issues

f1133c9

Refactor to better align with SDKs, better integration tests

02dfed0

Maybe fix windows

776478e

Doc updates, create token server helper (future example)

59b9446

Cleaner literal token source

f397a5b

Drop ConnectionDetails (doesn't align well with other SDKs/practical …

2e7d783

…usage)

Use nlohmann/json instead of bespoke impl

d5ad498

Add token_source_tester using LIVEKIT_SANDBOX_ID env var.

523c613

Manual ad-hoc helper for exercising token sources locally; the sandbox ID must be supplied via LIVEKIT_SANDBOX_ID rather than hard-coded in source. Co-authored-by: Cursor <cursoragent@cursor.com>

Try live token in CI

90d262f

alan-george-lk force-pushed the feature/token_source_api branch from 763d27e to 90d262f Compare June 22, 2026 19:52

Lots of testing changes

2229729

alan-george-lk added 7 commits June 23, 2026 10:34

Clean doc comment

1011803

Try new action

4bd9e42

Maybe fix windows build

1f13ef0

Try latest token-server-action

32dcfa8

Use tagged action

90e9d0d

Merge branch 'main' of github.com:livekit/client-sdk-cpp into feature…

c53e172

…/token_source_api

PR review, more docs, additional integration tests around tokens

acaecb6

xianshijing-lk approved these changes Jun 25, 2026

View reviewed changes

PR cleanup

7c77d8d

alan-george-lk requested a review from 1egoman June 25, 2026 17:57

Merge branch 'main' of github.com:livekit/client-sdk-cpp into feature…

e582359

…/token_source_api

1egoman reviewed Jun 25, 2026

View reviewed changes

alan-george-lk added 4 commits June 25, 2026 13:14

Some PR feedback

737c86e

Additional PR feedback

e39e55e

Additional PR feedback

e9d3ff0

Additional cleaup

bbf6a41

alan-george-lk marked this pull request as ready for review June 26, 2026 14:32

alan-george-lk requested a review from stephen-derosa as a code owner June 26, 2026 14:32

Doc update

0acfc57

		@@ -0,0 +1,109 @@
		# Token lifecycle

		Succinct reference for how join credentials and in-session refresh interact in

Uh oh!

Conversation

alan-george-lk commented Jun 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Summary

Server Token Refresh

New Dependencies

Testing

Token Server Action

Sandbox Token

Uh oh!

xianshijing-lk left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

MaxHeimbrock commented Jun 23, 2026

Uh oh!

xianshijing-lk left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

alan-george-lk Jun 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

alan-george-lk Jun 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

1egoman left a comment

Choose a reason for hiding this comment

Uh oh!

1egoman Jun 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

1egoman Jun 25, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

alan-george-lk Jun 26, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

alan-george-lk commented Jun 18, 2026 •

edited

Loading

alan-george-lk Jun 25, 2026 •

edited

Loading

alan-george-lk Jun 25, 2026 •

edited

Loading

1egoman Jun 25, 2026 •

edited

Loading

1egoman Jun 25, 2026 •

edited

Loading

alan-george-lk Jun 26, 2026 •

edited

Loading

alan-george-lk Jun 26, 2026 •

edited

Loading