Skip to content

Lightpanda: the headless browser designed for AI and automation

agent: map models_dev_id and filter tool-call models

Adrià Arrufat01a7b091db5c
diff --git a/build.zig.zon b/build.zig.zon
index ae87bd301..2f92925df 100644
--- a/build.zig.zon
+++ b/build.zig.zon
@@ -36,8 +36,8 @@
.hash = "sqlite3-3.53.2-DMxLWuAOAAA_Px0arJOIOaP4AKEu5prbsQgPMA35W1zz",
},
.zenai = .{
- .url = "git+https://github.com/lightpanda-io/zenai.git?ref=codex#69673248321adeee537dfedcf9c13eed18928c30",
- .hash = "zenai-0.0.0-iOY_VP0GBgCClHsRHVmHvSVfYVT33yty_Lb-Qafpb2Gm",
+ .url = "git+https://github.com/lightpanda-io/zenai.git#2fb52c9ef814b6b57596692477862d5183969d7c",
+ .hash = "zenai-0.0.0-iOY_VOb-BQCiU8PDX1qxvpJm-AQHLGFOp_DN6xIV5iJx",
},
.isocline = .{
.url = "git+https://github.com/arrufat/isocline?ref=lightpanda#832a9fe25f5f4458fcc47b5acc7c21db669c2f47",
diff --git a/src/agent/Agent.zig b/src/agent/Agent.zig
index 13201c207..887e029cb 100644
--- a/src/agent/Agent.zig
+++ b/src/agent/Agent.zig
@@ -1941,8 +1941,8 @@ pub fn listModels(allocator: std.mem.Allocator, opts: Config.Agent) !void {
defer arena.deinit();
// A subscription provider can't list models via the provider API; use the
// free models.dev catalog (uncached here — this is a one-shot CLI command).
- if (auth.descriptorFor(llm.provider) != null) {
- const sub_ids = models_dev.modelIds(arena.allocator(), @tagName(llm.provider), null);
+ if (auth.descriptorFor(llm.provider)) |desc| {
+ const sub_ids = models_dev.modelIds(arena.allocator(), desc.models_dev_id, null);
var stdout_sub = std.Io.File.stdout().writerStreaming(lp.io, &.{});
const ws = &stdout_sub.interface;
for (sub_ids) |id| try ws.print("{s}\n", .{id});
@@ -2026,8 +2026,8 @@ fn completionModels(context: *anyopaque, _: std.mem.Allocator) []const []const u
const arena = self.model_completion_arena.allocator();
// A subscription provider can't hit the provider's `/models` endpoint, so
// list from the free, unauthenticated models.dev catalog instead.
- const ids = if (auth.descriptorFor(llm.provider) != null)
- models_dev.modelIds(arena, @tagName(llm.provider), self.app_dir)
+ const ids = if (auth.descriptorFor(llm.provider)) |desc|
+ models_dev.modelIds(arena, desc.models_dev_id, self.app_dir)
else
zenai.provider.listChatModelIds(
lp.io,
diff --git a/src/agent/auth/auth.zig b/src/agent/auth/auth.zig
index 52ad92152..8d994f24b 100644
--- a/src/agent/auth/auth.zig
+++ b/src/agent/auth/auth.zig
@@ -73,6 +73,9 @@ pub const Descriptor = struct {
provider: Config.AiProvider,
/// Key in the on-disk store (`auth.json`).
id: []const u8,
+ /// Provider key in the models.dev catalog (subscription backends have no
+ /// entry of their own; codex's models are listed under "openai").
+ models_dev_id: []const u8,
/// Human label for the credential, e.g. "ChatGPT subscription".
label: []const u8,
client_id: []const u8,
diff --git a/src/agent/auth/codex.zig b/src/agent/auth/codex.zig
index e09f0ac3c..089aac379 100644
--- a/src/agent/auth/codex.zig
+++ b/src/agent/auth/codex.zig
@@ -46,6 +46,7 @@ const cancel_slice_ms: u64 = 200;
pub const descriptor: auth.Descriptor = .{
.provider = .codex,
.id = "codex",
+ .models_dev_id = "openai",
.label = "ChatGPT subscription",
.client_id = client_id,
.scope = scope,
diff --git a/src/agent/auth/models_dev.zig b/src/agent/auth/models_dev.zig
index 2a44f4d2d..30efef3c9 100644
--- a/src/agent/auth/models_dev.zig
+++ b/src/agent/auth/models_dev.zig
@@ -74,23 +74,43 @@ fn writeCache(app_dir: []const u8, provider_id: []const u8, ids: []const []const
try std.Io.Dir.cwd().writeFile(lp.io, .{ .sub_path = path, .data = buf.written() });
}
-/// Model-id keys of `catalog[provider_id].models`. `ignore_unknown_fields` skips
-/// every provider and model's metadata, so only the id strings are built — no
-/// Value tree for the multi-MB payload. The parse runs in a scoped arena; the
-/// ids are duped into `arena`.
+/// Model-id keys of `catalog[provider_id].models`, filtered to tool-call-capable
+/// models (the agent needs tools; this also drops embedding/image entries).
+/// `ignore_unknown_fields` skips the rest of the metadata, so no Value tree is
+/// built for the multi-MB payload. The parse runs in a scoped arena; the ids
+/// are duped into `arena`.
fn parseProviderModels(arena: std.mem.Allocator, catalog: []const u8, provider_id: []const u8) ![]const []const u8 {
var parse_arena: std.heap.ArenaAllocator = .init(arena);
defer parse_arena.deinit();
- const Empty = struct {};
- const Provider = struct { models: std.json.ArrayHashMap(Empty) = .{} };
+ const Model = struct { tool_call: bool = false };
+ const Provider = struct { models: std.json.ArrayHashMap(Model) = .{} };
const parsed = try std.json.parseFromSliceLeaky(std.json.ArrayHashMap(Provider), parse_arena.allocator(), catalog, .{ .ignore_unknown_fields = true });
const provider = parsed.map.get(provider_id) orelse return error.ProviderMissing;
- const keys = provider.models.map.keys();
- const ids = try arena.alloc([]const u8, keys.len);
- for (keys, ids) |k, *dst| dst.* = try arena.dupe(u8, k);
- return ids;
+ var ids: std.ArrayList([]const u8) = .empty;
+ var it = provider.models.map.iterator();
+ while (it.next()) |entry| {
+ if (!entry.value_ptr.tool_call) continue;
+ try ids.append(arena, try arena.dupe(u8, entry.key_ptr.*));
+ }
+ return ids.toOwnedSlice(arena);
+}
+
+test "parseProviderModels filters to tool-call models" {
+ var arena: std.heap.ArenaAllocator = .init(std.testing.allocator);
+ defer arena.deinit();
+ const catalog =
+ \\{"openai":{"models":{
+ \\ "gpt-5.5":{"tool_call":true},
+ \\ "text-embedding-3-small":{"tool_call":false},
+ \\ "gpt-image-1":{}
+ \\}}}
+ ;
+ const ids = try parseProviderModels(arena.allocator(), catalog, "openai");
+ try std.testing.expectEqual(1, ids.len);
+ try std.testing.expectEqualStrings("gpt-5.5", ids[0]);
+ try std.testing.expectError(error.ProviderMissing, parseProviderModels(arena.allocator(), catalog, "nope"));
}
fn fetch(arena: std.mem.Allocator) ![]u8 {