diff --git a/Gemfile.lock b/Gemfile.lock
index d9908f5e1..3494b4489 100644
--- a/Gemfile.lock
+++ b/Gemfile.lock
@@ -275,14 +275,16 @@ GEM
     fugit (1.11.1)
       et-orbi (~> 1, >= 1.2.11)
       raabro (~> 1.4)
-    gapic-common (0.20.0)
+    gapic-common (1.0.0)
       faraday (>= 1.9, < 3.a)
       faraday-retry (>= 1.0, < 3.a)
-      google-protobuf (~> 3.14)
-      googleapis-common-protos (>= 1.3.12, < 2.a)
-      googleapis-common-protos-types (>= 1.3.1, < 2.a)
-      googleauth (~> 1.0)
-      grpc (~> 1.36)
+      google-cloud-env (~> 2.2)
+      google-logging-utils (~> 0.1)
+      google-protobuf (>= 3.25, < 5.a)
+      googleapis-common-protos (~> 1.6)
+      googleapis-common-protos-types (~> 1.15)
+      googleauth (~> 1.12)
+      grpc (~> 1.66)
     geocoder (1.8.1)
     gli (2.22.2)
       ostruct
@@ -327,6 +329,13 @@ GEM
     google-cloud-translate-v3 (0.10.0)
       gapic-common (>= 0.20.0, < 2.a)
       google-cloud-errors (~> 1.0)
+    google-cloud-vision (2.0.2)
+      google-cloud-core (~> 1.6)
+      google-cloud-vision-v1 (~> 1.2)
+    google-cloud-vision-v1 (1.3.0)
+      gapic-common (~> 1.0)
+      google-cloud-errors (~> 1.0)
+    google-logging-utils (0.2.0)
     google-protobuf (3.25.7)
     googleapis-common-protos (1.6.0)
       google-protobuf (>= 3.18, < 5.a)
@@ -334,9 +343,10 @@ GEM
       grpc (~> 1.41)
     googleapis-common-protos-types (1.20.0)
       google-protobuf (>= 3.18, < 5.a)
-    googleauth (1.11.2)
+    googleauth (1.14.0)
       faraday (>= 1.0, < 3.a)
-      google-cloud-env (~> 2.1)
+      google-cloud-env (~> 2.2)
+      google-logging-utils (~> 0.1)
       jwt (>= 1.4, < 3.0)
       multi_json (~> 1.11)
       os (>= 0.9, < 2.0)
@@ -381,6 +391,7 @@ GEM
     httpclient (2.8.3)
     i18n (1.14.7)
       concurrent-ruby (~> 1.0)
+    iconv (1.1.0)
     image_processing (1.12.2)
       mini_magick (>= 4.9.5, < 5)
       ruby-vips (>= 2.0.17, < 3)
@@ -550,6 +561,9 @@ GEM
     parser (3.3.8.0)
       ast (~> 2.4.1)
       racc
+    pdftoimage (0.2.1)
+      iconv (~> 1.0)
+      shellwords (~> 0.2.2)
     pg (1.5.3)
     pg_search (2.3.6)
       activerecord (>= 5.2)
@@ -672,6 +686,7 @@ GEM
     rspec-support (3.13.1)
     rspec_junit_formatter (0.6.0)
       rspec-core (>= 2, < 4, != 2.12.0)
+    rtesseract (3.1.4)
     rubocop (1.75.6)
       json (~> 2.3)
       language_server-protocol (~> 3.17.0.2)
@@ -747,6 +762,7 @@ GEM
       sentry-ruby (~> 5.19.0)
       sidekiq (>= 3.0)
     sexp_processor (4.17.0)
+    shellwords (0.2.2)
     shopify_api (14.8.0)
       activesupport
       concurrent-ruby
@@ -927,6 +943,7 @@ DEPENDENCIES
   google-cloud-dialogflow-v2 (>= 0.24.0)
   google-cloud-storage (>= 1.48.0)
   google-cloud-translate-v3 (>= 0.7.0)
+  google-cloud-vision
   groupdate
   grpc
   haikunator
@@ -950,6 +967,7 @@ DEPENDENCIES
   lograge (~> 0.14.0)
   maxminddb
   meta_request (>= 0.8.3)
+  mini_magick
   mock_redis
   neighbor
   net-smtp (~> 0.3.4)
@@ -959,6 +977,7 @@ DEPENDENCIES
   omniauth-google-oauth2 (>= 1.1.3)
   omniauth-oauth2
   omniauth-rails_csrf_protection (~> 1.0, >= 1.0.2)
+  pdftoimage
   pg
   pg_search
   pgvector
@@ -978,6 +997,7 @@ DEPENDENCIES
   reverse_markdown
   rspec-rails (>= 6.1.5)
   rspec_junit_formatter
+  rtesseract
   rubocop
   rubocop-factory_bot
   rubocop-performance
@@ -1020,4 +1040,4 @@ RUBY VERSION
    ruby 3.4.4p34
 
 BUNDLED WITH
-   2.5.16
+   2.6.7
diff --git a/db/migrate/20250623120000_add_document_type_and_file_to_captain_documents.rb b/db/migrate/20250623120000_add_document_type_and_file_to_captain_documents.rb
new file mode 100644
index 000000000..12769e704
--- /dev/null
+++ b/db/migrate/20250623120000_add_document_type_and_file_to_captain_documents.rb
@@ -0,0 +1,6 @@
+class AddDocumentTypeAndFileToCaptainDocuments < ActiveRecord::Migration[7.0]
+  def change
+    add_column :captain_documents, :document_type, :integer, default: 0, null: false
+    add_index :captain_documents, :document_type
+  end
+end
\ No newline at end of file
diff --git a/db/schema.rb b/db/schema.rb
index d065a8ff4..3f5839eaa 100644
--- a/db/schema.rb
+++ b/db/schema.rb
@@ -10,7 +10,7 @@
 #
 # It's strongly recommended that you check this file into your version control system.
 
-ActiveRecord::Schema[7.1].define(version: 2025_05_23_031839) do
+ActiveRecord::Schema[7.1].define(version: 2025_06_23_120000) do
   # These extensions should be enabled to support this database
   enable_extension "pg_stat_statements"
   enable_extension "pg_trgm"
@@ -289,9 +289,11 @@ ActiveRecord::Schema[7.1].define(version: 2025_05_23_031839) do
     t.datetime "created_at", null: false
     t.datetime "updated_at", null: false
     t.integer "status", default: 0, null: false
+    t.integer "document_type", default: 0, null: false
     t.index ["account_id"], name: "index_captain_documents_on_account_id"
     t.index ["assistant_id", "external_link"], name: "index_captain_documents_on_assistant_id_and_external_link", unique: true
     t.index ["assistant_id"], name: "index_captain_documents_on_assistant_id"
+    t.index ["document_type"], name: "index_captain_documents_on_document_type"
     t.index ["status"], name: "index_captain_documents_on_status"
   end
 
@@ -907,7 +909,7 @@ ActiveRecord::Schema[7.1].define(version: 2025_05_23_031839) do
     t.text "header_text"
     t.datetime "created_at", null: false
     t.datetime "updated_at", null: false
-    t.jsonb "config", default: {"allowed_locales"=>["en"]}
+    t.jsonb "config", default: {"allowed_locales" => ["en"]}
     t.boolean "archived", default: false
     t.bigint "channel_web_widget_id"
     t.index ["channel_web_widget_id"], name: "index_portals_on_channel_web_widget_id"
diff --git a/docker/Dockerfile b/docker/Dockerfile
index 2d753337d..781440aa2 100644
--- a/docker/Dockerfile
+++ b/docker/Dockerfile
@@ -34,6 +34,8 @@ RUN apk update && apk add --no-cache \
   git \
   curl \
   xz \
+  poppler-utils \
+  tesseract-ocr \
   && mkdir -p /var/app \
   && gem install bundler
 
@@ -128,6 +130,8 @@ RUN apk update && apk add --no-cache \
   imagemagick \
   git \
   vips \
+  poppler-utils \
+  tesseract-ocr \
   && gem install bundler
 
 COPY --from=node /usr/local/bin/node /usr/local/bin/
diff --git a/enterprise/app/models/captain/document.rb b/enterprise/app/models/captain/document.rb
index 6ed8d6e8a..3e5b1e43e 100644
--- a/enterprise/app/models/captain/document.rb
+++ b/enterprise/app/models/captain/document.rb
@@ -4,6 +4,7 @@
 #
 #  id            :bigint           not null, primary key
 #  content       :text
+#  document_type :integer          default("url"), not null
 #  external_link :string           not null
 #  name          :string
 #  status        :integer          default("in_progress"), not null
@@ -17,6 +18,7 @@
 #  index_captain_documents_on_account_id                      (account_id)
 #  index_captain_documents_on_assistant_id                    (assistant_id)
 #  index_captain_documents_on_assistant_id_and_external_link  (assistant_id,external_link) UNIQUE
+#  index_captain_documents_on_document_type                   (document_type)
 #  index_captain_documents_on_status                          (status)
 #
 class Captain::Document < ApplicationRecord
@@ -53,6 +55,14 @@ class Captain::Document < ApplicationRecord
   scope :for_account, ->(account_id) { where(account_id: account_id) }
   scope :for_assistant, ->(assistant_id) { where(assistant_id: assistant_id) }
 
+  def url_document?
+    document_type == 'url'
+  end
+
+  def pdf_document?
+    document_type == 'pdf'
+  end
+
   private
 
   def enqueue_crawl_job
@@ -65,14 +75,6 @@ class Captain::Document < ApplicationRecord
     end
   end
 
-  def url_document?
-    document_type == 'url'
-  end
-
-  def pdf_document?
-    document_type == 'pdf'
-  end
-
   def detect_document_type
     return unless external_link.present?