llm.rb 15.2.2 → 15.4.0

data/lib/llm/provider.rb

@@ -255,6 +255,22 @@ class LLM::Provider
     raise NotImplementedError
   end
 
+  ##
+  # Returns the number of times a rate-limited
+  # request is retried before the error is raised.
+  #
+  # Each retry sleeps a growing interval, so a budget
+  # that is exhausted surfaces the rate-limit error
+  # instead of blocking indefinitely. A provider can
+  # return a different budget when its API recovers
+  # from rate limits more slowly.
+  #
+  # @see LLM::Agent#retry_budget
+  # @return [Integer]
+  def retry_budget
+    5
+  end
+
   ##
   # Returns an object that can generate a JSON schema
   # @return [LLM::Schema]
@@ -263,11 +279,21 @@ class LLM::Provider
   end
 
   ##
-  # Add one or more headers to all requests
-  # @example
+  # Add one or more headers to all requests, or scope them to a block.
+  #
+  # Without a block the headers are merged into the provider's defaults
+  # permanently. With a block the headers apply only to the current fiber,
+  # for the duration of the block, then the previous headers are restored.
+  # This is useful for a header that varies per context, such as
+  # OpenRouter's `x-session-id`.
+  # @example Permanent
   #   llm = LLM.openai(key: ENV["KEY"])
   #   llm.with("OpenAI-Organization" => ENV["ORG"])
   #   llm.with("OpenAI-Project" => ENV["PROJECT"])
+  # @example Scoped
+  #   llm.with("x-session-id" => ctx.id) do
+  #     llm.complete("hello")
+  #   end
   # @param [Hash<String,String>] headers
   #  One or more headers
   # @note
@@ -275,12 +301,26 @@ class LLM::Provider
   #  provided via the `headers:` keyword argument,
   #  or provided directly as a Hash without the
   #  `headers:` key namespace.
+  # @yield
   # @return [LLM::Provider]
-  #  Returns self
-  def with(**headers)
+  #  Returns self without a block, the block's value with one
+  def with(**headers, &block)
     headers = headers.merge(headers.delete(:headers) || {})
-    lock do
-      tap { @headers.merge!(headers) }
+    if block
+      wm = weakmaps.header
+      previous = wm[self]
+      wm[self] = (previous || {}).merge(headers)
+      block.call
+    else
+      lock { tap { @headers.merge!(headers) } }
+    end
+  ensure
+    if block
+      if previous.nil?
+        wm.respond_to?(:delete) ? wm.delete(self) : wm[self] = nil
+      else
+        wm[self] = previous
+      end
     end
   end
 
@@ -349,7 +389,7 @@ class LLM::Provider
   # @return [LLM::Tracer]
   #  Returns the current scoped tracer override or provider default tracer
   def tracer
-    weakmap[self] || @tracer || LLM::Tracer::Null.new(self)
+    weakmaps.tracer[self] || @tracer || LLM::Tracer::Null.new(self)
   end
 
   ##
@@ -378,17 +418,18 @@ class LLM::Provider
   # @yield
   # @return [Object]
   def with_tracer(tracer)
-    had_override = weakmap.key?(self)
-    previous = weakmap[self]
-    weakmap[self] = tracer || LLM::Tracer::Null.new(self)
+    wm = weakmaps.tracer
+    had_override = wm.key?(self)
+    previous = wm[self]
+    wm[self] = tracer || LLM::Tracer::Null.new(self)
     yield
   ensure
     if had_override
-      weakmap[self] = previous
-    elsif weakmap.respond_to?(:delete)
-      weakmap.delete(self)
+      wm[self] = previous
+    elsif wm.respond_to?(:delete)
+      wm.delete(self)
     else
-      weakmap[self] = nil
+      wm[self] = nil
     end
   end
 
@@ -504,8 +545,25 @@ class LLM::Provider
   end
 
   ##
+  # Returns the temporary headers set by {#with}
+  # for the current fiber.
   # @api private
-  def weakmap
-    thread[:"llm.provider.weakmap"] ||= ObjectSpace::WeakMap.new
+  # @return [Hash]
+  def temporary_headers
+    weakmaps.header[self] || {}
+  end
+
+  ##
+  # @return [LLM::Object]
+  def weakmaps
+    if thread["llm.#{name}.weakmaps"]
+      thread["llm.#{name}.weakmaps"]
+    else
+      weakmaps = LLM::Object.from(
+        tracer: ObjectSpace::WeakMap.new,
+        header: ObjectSpace::WeakMap.new
+      )
+      thread["llm.#{name}.weakmaps"] = weakmaps
+    end
   end
 end

data/lib/llm/providers/alibaba.rb

@@ -89,5 +89,20 @@ module LLM
     def default_model
       "deepseek-v4-flash-0731"
     end
+
+    ##
+    # Returns the number of times a rate-limited
+    # request is retried.
+    #
+    # Alibaba (token plan) will frequently issue
+    # rate limits or time outs that it recovers from.
+    # The higher retry count is to account for scenarios
+    # where it takes longer than expected to recover.
+    #
+    # @see LLM::Provider#retry_budget
+    # @return [Integer]
+    def retry_budget
+      8
+    end
   end
 end

data/lib/llm/providers/anthropic/error_handler.rb

@@ -21,10 +21,13 @@ class LLM::Anthropic
     #  The span
     # @param [LLM::Transport::Response, Net::HTTPResponse] res
     #  The response from the server
+    # @param [String] request_id
+    #  The id of the request that failed
     # @return [LLM::Anthropic::ErrorHandler]
-    def initialize(tracer, span, res)
+    def initialize(tracer, span, res, request_id)
       @tracer = tracer
       @span = span
+      @request_id = request_id
       @res = LLM::Transport::Response.from(res)
     end
 
@@ -33,7 +36,7 @@ class LLM::Anthropic
     #  Raises a subclass of {LLM::Error LLM::Error}
     def raise_error!
       ex = error
-      @tracer.on_request_error(ex:, span:)
+      @tracer.on_request_error(ex:, span:, request_id: @request_id)
     ensure
       raise(ex) if ex
     end

data/lib/llm/providers/anthropic/files.rb

@@ -38,9 +38,9 @@ class LLM::Anthropic
     def all(**params)
       query = URI.encode_www_form(params)
       req = LLM::Transport::Request.get("/v1/files?#{query}", headers)
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = ResponseAdapter.adapt(res, type: :enumerable)
-      tracer.on_request_finish(operation: "request", res:, span:)
+      tracer.on_request_finish(operation: "request", res:, span:, request_id:)
       res
     end
 
@@ -59,9 +59,9 @@ class LLM::Anthropic
       req = LLM::Transport::Request.post("/v1/files", headers)
       req["content-type"] = multi.content_type
       transport.set_body_stream(req, multi.body)
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = ResponseAdapter.adapt(res, type: :file)
-      tracer.on_request_finish(operation: "request", res:, span:)
+      tracer.on_request_finish(operation: "request", res:, span:, request_id:)
       res
     end
 
@@ -80,9 +80,9 @@ class LLM::Anthropic
       file_id = file.respond_to?(:id) ? file.id : file
       query = URI.encode_www_form(params)
       req = LLM::Transport::Request.get("/v1/files/#{file_id}?#{query}", headers)
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = ResponseAdapter.adapt(res, type: :file)
-      tracer.on_request_finish(operation: "request", res:, span:)
+      tracer.on_request_finish(operation: "request", res:, span:, request_id:)
       res
     end
 
@@ -101,9 +101,9 @@ class LLM::Anthropic
       query = URI.encode_www_form(params)
       file_id = file.respond_to?(:id) ? file.id : file
       req = LLM::Transport::Request.get("/v1/files/#{file_id}?#{query}", headers)
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = ResponseAdapter.adapt(res, type: :file)
-      tracer.on_request_finish(operation: "request", res:, span:)
+      tracer.on_request_finish(operation: "request", res:, span:, request_id:)
       res
     end
     alias_method :retrieve_metadata, :get_metadata
@@ -121,9 +121,9 @@ class LLM::Anthropic
     def delete(file:)
       file_id = file.respond_to?(:id) ? file.id : file
       req = LLM::Transport::Request.delete("/v1/files/#{file_id}", headers)
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = LLM::Response.new(res)
-      tracer.on_request_finish(operation: "request", res:, span:)
+      tracer.on_request_finish(operation: "request", res:, span:, request_id:)
       res
     end
 
@@ -147,9 +147,9 @@ class LLM::Anthropic
       file_id = file.respond_to?(:id) ? file.id : file
       req = LLM::Transport::Request.get("/v1/files/#{file_id}/content?#{query}", headers)
       io = StringIO.new("".b)
-      res, span, tracer = execute(request: req, operation: "request") { |res| res.read_body { |chunk| io << chunk } }
+      res, span, tracer, request_id = execute(request: req, operation: "request") { |res| res.read_body { |chunk| io << chunk } }
       res = LLM::Response.new(res).tap { _1.define_singleton_method(:file) { io } }
-      tracer.on_request_finish(operation: "request", res:, span:)
+      tracer.on_request_finish(operation: "request", res:, span:, request_id:)
       res
     end
 

data/lib/llm/providers/anthropic/models.rb

@@ -40,9 +40,9 @@ class LLM::Anthropic
     def all(**params)
       query = URI.encode_www_form(params)
       req = LLM::Transport::Request.get("/v1/models?#{query}", headers)
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = ResponseAdapter.adapt(res, type: :models)
-      tracer.on_request_finish(operation: "request", res:, span:)
+      tracer.on_request_finish(operation: "request", res:, span:, request_id:)
       res
     end
 

data/lib/llm/providers/anthropic.rb

@@ -52,10 +52,10 @@ module LLM
     def complete(prompt, params = {})
       params, stream, tools, role = normalize_complete_params(params)
       req = build_complete_request(prompt, params, role)
-      res, span, tracer = execute(request: req, stream: stream, operation: "chat", model: params[:model])
+      res, span, tracer, request_id = execute(request: req, stream: stream, operation: "chat", model: params[:model])
       res = ResponseAdapter.adapt(res, type: :completion)
         .extend(Module.new { define_method(:__tools__) { tools } })
-      tracer.on_request_finish(operation: "chat", model: params[:model], res:, span:)
+      tracer.on_request_finish(operation: "chat", model: params[:model], res:, span:, request_id:)
       res
     end
 
@@ -141,7 +141,9 @@ module LLM
 
     def headers
       lock do
-        (@headers || {}).merge(
+        (@headers || {})
+        .merge(temporary_headers)
+        .merge(
           "Content-Type" => "application/json",
           "x-api-key" => @key,
           "anthropic-version" => "2023-06-01",

data/lib/llm/providers/bedrock/error_handler.rb

@@ -23,9 +23,10 @@ class LLM::Bedrock
     # @param [Object, nil] span
     # @param [LLM::Transport::Response, Net::HTTPResponse] res
     # @return [LLM::Bedrock::ErrorHandler]
-    def initialize(tracer, span, res)
+    def initialize(tracer, span, res, request_id)
       @tracer = tracer
       @span = span
+      @request_id = request_id
       @res = LLM::Transport::Response.from(res)
     end
 
@@ -33,7 +34,7 @@ class LLM::Bedrock
     # @raise [LLM::Error]
     def raise_error!
       ex = error
-      @tracer.on_request_error(ex:, span:)
+      @tracer.on_request_error(ex:, span:, request_id: @request_id)
     ensure
       raise(ex) if ex
     end

data/lib/llm/providers/bedrock/models.rb

@@ -40,9 +40,10 @@ class LLM::Bedrock
     # @return [LLM::Response]
     def all(**params)
       host = credentials.host
+      request_id = SecureRandom.uuid_v7
       req = build_request(host, params)
       res = build_transport(host).request(req, owner: self)
-      handle_response(res)
+      handle_response(res, request_id)
     end
 
     private
@@ -69,9 +70,10 @@ class LLM::Bedrock
 
     ##
     # @param [LLM::Transport::Response, Net::HTTPResponse] res
+    # @param [String] request_id
     # @return [LLM::Response]
     # @raise [LLM::Error]
-    def handle_response(res)
+    def handle_response(res, request_id)
       res = LLM::Transport::Response.from(res)
       if res.success?
         res.body = LLM::Object.from(LLM.json.load(res.body || "{}"))
@@ -79,7 +81,7 @@ class LLM::Bedrock
       else
         body = +""
         res.read_body { body << _1 } if res.body.nil?
-        LLM::Bedrock::ErrorHandler.new(tracer, nil, res).raise_error!
+        LLM::Bedrock::ErrorHandler.new(tracer, nil, res, request_id).raise_error!
       end
     end
 

data/lib/llm/providers/bedrock.rb

@@ -89,10 +89,10 @@ module LLM
       tracer.set_request_metadata(user_input: extract_user_input(messages, fallback: prompt))
       sign!(req, body)
       model_id = model_id_for(req.path)
-      res, span, tracer = execute(request: req, stream:, operation: "chat", stream_parser:, model: model_id)
+      res, span, tracer, request_id = execute(request: req, stream:, operation: "chat", stream_parser:, model: model_id)
       res = ResponseAdapter.adapt(res, type: :completion)
         .extend(Module.new { define_method(:__tools__) { tools } })
-      tracer.on_request_finish(operation: "chat", model: model_id, res:, span:)
+      tracer.on_request_finish(operation: "chat", model: model_id, res:, span:, request_id:)
       res
     end
 
@@ -183,7 +183,9 @@ module LLM
 
     def headers
       lock do
-        (@headers || {}).merge("Content-Type" => "application/json")
+        (@headers || {})
+        .merge(temporary_headers)
+        .merge("Content-Type" => "application/json")
       end
     end
 

data/lib/llm/providers/deepinfra/audio.rb

@@ -23,9 +23,9 @@ class LLM::DeepInfra
       path = path("/v1/inference/#{model}", base_path: false)
       req = LLM::Transport::Request.post(path, headers)
       req.body = JSON.dump(params.merge(text: input))
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = ResponseAdapter.adapt LLM::Response.new(res), type: :audio
-      tracer.on_request_finish(operation: "request", model:, res:, span:)
+      tracer.on_request_finish(operation: "request", model:, res:, span:, request_id:)
       res
     end
 
@@ -45,9 +45,9 @@ class LLM::DeepInfra
       req = LLM::Transport::Request.post(path, headers)
       req["content-type"] = multi.content_type
       transport.set_body_stream(req, multi.body)
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = LLM::Response.new(res)
-      tracer.on_request_finish(operation: "request", model:, res:, span:)
+      tracer.on_request_finish(operation: "request", model:, res:, span:, request_id:)
       res
     end
 

data/lib/llm/providers/deepinfra/images.rb

@@ -44,9 +44,9 @@ class LLM::DeepInfra
       req = LLM::Transport::Request.post(path("/images/generations"), headers)
       params = {prompt:, model:, size:, n:, response_format:, quality:, style:}.compact
       req.body = LLM.json.dump(params)
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = LLM::OpenAI::ResponseAdapter.adapt(res, type: :image)
-      tracer.on_request_finish(operation: "request", model:, res:, span:)
+      tracer.on_request_finish(operation: "request", model:, res:, span:, request_id:)
       res
     end
 
@@ -75,9 +75,9 @@ class LLM::DeepInfra
       req = LLM::Transport::Request.post(path("/images/edits"), headers)
       req["content-type"] = multi.content_type
       transport.set_body_stream(req, multi.body)
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = LLM::OpenAI::ResponseAdapter.adapt(res, type: :image)
-      tracer.on_request_finish(operation: "request", model:, res:, span:)
+      tracer.on_request_finish(operation: "request", model:, res:, span:, request_id:)
       res
     end
 

data/lib/llm/providers/google/error_handler.rb

@@ -21,10 +21,13 @@ class LLM::Google
     #  The span
     # @param [LLM::Transport::Response, Net::HTTPResponse] res
     #  The response from the server
+    # @param [String] request_id
+    #  The id of the request that failed
     # @return [LLM::Google::ErrorHandler]
-    def initialize(tracer, span, res)
+    def initialize(tracer, span, res, request_id)
       @tracer = tracer
       @span = span
+      @request_id = request_id
       @res = LLM::Transport::Response.from(res)
     end
 
@@ -33,7 +36,7 @@ class LLM::Google
     #  Raises a subclass of {LLM::Error LLM::Error}
     def raise_error!
       ex = error
-      @tracer.on_request_error(ex:, span:)
+      @tracer.on_request_error(ex:, span:, request_id: @request_id)
     ensure
       raise(ex) if ex
     end

data/lib/llm/providers/google/files.rb

@@ -46,9 +46,9 @@ class LLM::Google
     def all(**params)
       query = URI.encode_www_form(params.merge!(key: key))
       req = LLM::Transport::Request.get("/v1beta/files?#{query}", headers)
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = ResponseAdapter.adapt(res, type: :files)
-      tracer.on_request_finish(operation: "request", res:, span:)
+      tracer.on_request_finish(operation: "request", res:, span:, request_id:)
       res
     end
 
@@ -70,9 +70,9 @@ class LLM::Google
       req["X-Goog-Upload-Command"] = "upload, finalize"
       file.with_io do |io|
         transport.set_body_stream(req, io)
-        res, span, tracer = execute(request: req, operation: "request")
+        res, span, tracer, request_id = execute(request: req, operation: "request")
         res = ResponseAdapter.adapt(res, type: :file)
-        tracer.on_request_finish(operation: "request", res:, span:)
+        tracer.on_request_finish(operation: "request", res:, span:, request_id:)
         res
       end
     end
@@ -92,9 +92,9 @@ class LLM::Google
       file_id = file.respond_to?(:name) ? file.name : file.to_s
       query = URI.encode_www_form(params.merge!(key: key))
       req = LLM::Transport::Request.get("/v1beta/#{file_id}?#{query}", headers)
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = ResponseAdapter.adapt(res, type: :file)
-      tracer.on_request_finish(operation: "request", res:, span:)
+      tracer.on_request_finish(operation: "request", res:, span:, request_id:)
       res
     end
 
@@ -112,9 +112,9 @@ class LLM::Google
       file_id = file.respond_to?(:name) ? file.name : file.to_s
       query = URI.encode_www_form(params.merge!(key: key))
       req = LLM::Transport::Request.delete("/v1beta/#{file_id}?#{query}", headers)
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = LLM::Response.new(res)
-      tracer.on_request_finish(operation: "request", res:, span:)
+      tracer.on_request_finish(operation: "request", res:, span:, request_id:)
       res
     end
 
@@ -134,9 +134,9 @@ class LLM::Google
       req["X-Goog-Upload-Header-Content-Length"] = file.bytesize
       req["X-Goog-Upload-Header-Content-Type"] = file.mime_type
       req.body = LLM.json.dump({file: {display_name: File.basename(file.path)}})
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       trace_res = LLM::Response.new(res)
-      tracer.on_request_finish(operation: "request", res: trace_res, span:)
+      tracer.on_request_finish(operation: "request", res: trace_res, span:, request_id:)
       res["x-goog-upload-url"]
     end
 

data/lib/llm/providers/google/images.rb

@@ -51,9 +51,9 @@ class LLM::Google
         instances: [{prompt:}]
       })
       req.body = body
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = ResponseAdapter.adapt(res, type: :image)
-      tracer.on_request_finish(operation: "request", model:, res:, span:)
+      tracer.on_request_finish(operation: "request", model:, res:, span:, request_id:)
       res
     end
 

data/lib/llm/providers/google/models.rb

@@ -40,9 +40,9 @@ class LLM::Google
     def all(**params)
       query = URI.encode_www_form(params.merge!(key: key))
       req = LLM::Transport::Request.get("/v1beta/models?#{query}", headers)
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = ResponseAdapter.adapt(res, type: :models)
-      tracer.on_request_finish(operation: "request", res:, span:)
+      tracer.on_request_finish(operation: "request", res:, span:, request_id:)
       res
     end
 

data/lib/llm/providers/google.rb

@@ -58,9 +58,9 @@ module LLM
       path = ["/v1beta/models/#{model}", "embedContent?key=#{@key}"].join(":")
       req = LLM::Transport::Request.post(path, headers)
       req.body = LLM.json.dump({content: {parts: [{text: input}]}})
-      res, span, tracer = execute(request: req, operation: "embeddings", model:)
+      res, span, tracer, request_id = execute(request: req, operation: "embeddings", model:)
       res = ResponseAdapter.adapt(res, type: :embedding)
-      tracer.on_request_finish(operation: "embeddings", model:, res:, span:)
+      tracer.on_request_finish(operation: "embeddings", model:, res:, span:, request_id:)
       res
     end
 
@@ -77,10 +77,10 @@ module LLM
     def complete(prompt, params = {})
       params, stream, tools, role, model = normalize_complete_params(params)
       req = build_complete_request(prompt, params, role, model, stream)
-      res, span, tracer = execute(request: req, stream: stream, operation: "chat", model:)
+      res, span, tracer, request_id = execute(request: req, stream: stream, operation: "chat", model:)
       res = ResponseAdapter.adapt(res, type: :completion)
         .extend(Module.new { define_method(:__tools__) { tools } })
-      tracer.on_request_finish(operation: "chat", model:, res:, span:)
+      tracer.on_request_finish(operation: "chat", model:, res:, span:, request_id:)
       res
     end
 
@@ -186,9 +186,9 @@ module LLM
 
     def headers
       lock do
-        (@headers || {}).merge(
-          "Content-Type" => "application/json"
-        )
+        (@headers || {})
+        .merge(temporary_headers)
+        .merge("Content-Type" => "application/json")
       end
     end
 

data/lib/llm/providers/mistral.rb

@@ -136,7 +136,9 @@ module LLM
     # @api private
     def headers
       lock do
-        (@headers || {}).merge(
+        (@headers || {})
+        .merge(temporary_headers)
+        .merge(
           "Authorization" => "Bearer #{@key}",
           "Content-Type" => "application/json"
         )

data/lib/llm/providers/ollama/error_handler.rb

@@ -21,10 +21,13 @@ class LLM::Ollama
     #  The span
     # @param [LLM::Transport::Response, Net::HTTPResponse] res
     #  The response from the server
+    # @param [String] request_id
+    #  The id of the request that failed
     # @return [LLM::Ollama::ErrorHandler]
-    def initialize(tracer, span, res)
+    def initialize(tracer, span, res, request_id)
       @tracer = tracer
       @span = span
+      @request_id = request_id
       @res = LLM::Transport::Response.from(res)
     end
 
@@ -33,7 +36,7 @@ class LLM::Ollama
     #  Raises a subclass of {LLM::Error LLM::Error}
     def raise_error!
       ex = error
-      @tracer.on_request_error(ex:, span:)
+      @tracer.on_request_error(ex:, span:, request_id: @request_id)
     ensure
       raise(ex) if ex
     end

data/lib/llm/providers/ollama/models.rb

@@ -41,9 +41,9 @@ class LLM::Ollama
     def all(**params)
       query = URI.encode_www_form(params)
       req = LLM::Transport::Request.get("/api/tags?#{query}", headers)
-      res, span, tracer = execute(request: req, operation: "request")
+      res, span, tracer, request_id = execute(request: req, operation: "request")
       res = ResponseAdapter.adapt(res, type: :models)
-      tracer.on_request_finish(operation: "request", res:, span:)
+      tracer.on_request_finish(operation: "request", res:, span:, request_id:)
       res
     end
 

data/lib/llm/providers/ollama.rb

@@ -50,9 +50,9 @@ module LLM
       params   = {model:}.merge!(params)
       req      = LLM::Transport::Request.post("/v1/embeddings", headers)
       req.body = LLM.json.dump({input:}.merge!(params))
-      res, span, tracer = execute(request: req, operation: "embeddings", model:)
+      res, span, tracer, request_id = execute(request: req, operation: "embeddings", model:)
       res = ResponseAdapter.adapt(res, type: :embedding)
-      tracer.on_request_finish(operation: "embeddings", model:, res:, span:)
+      tracer.on_request_finish(operation: "embeddings", model:, res:, span:, request_id:)
       res
     end
 
@@ -69,10 +69,10 @@ module LLM
     def complete(prompt, params = {})
       params, stream, tools, role = normalize_complete_params(params)
       req = build_complete_request(prompt, params, role)
-      res, span, tracer = execute(request: req, stream: stream, operation: "chat", model: params[:model])
+      res, span, tracer, request_id = execute(request: req, stream: stream, operation: "chat", model: params[:model])
       res = ResponseAdapter.adapt(res, type: :completion)
         .extend(Module.new { define_method(:__tools__) { tools } })
-      tracer.on_request_finish(operation: "chat", model: params[:model], res:, span:)
+      tracer.on_request_finish(operation: "chat", model: params[:model], res:, span:, request_id:)
       res
     end
 
@@ -113,7 +113,9 @@ module LLM
 
     def headers
       lock do
-        (@headers || {}).merge(
+        (@headers || {})
+        .merge(temporary_headers)
+        .merge(
           "Content-Type" => "application/json",
           "Authorization" => "Bearer #{@key}"
         )