diff --git a/app/models/stream_recording.rb b/app/models/stream_recording.rb
index ad825158a..4d8a5357d 100644
--- a/app/models/stream_recording.rb
+++ b/app/models/stream_recording.rb
@@ -9,6 +9,7 @@ class StreamRecording < ApplicationRecord
   serialize :record_hours, coder: JSON
 
   belongs_to :podcast, -> { with_deleted }, touch: true, optional: true
+  has_many :stream_resources, -> { order("start_at DESC") }, dependent: :destroy
 
   scope :active, ->(now = Time.now) { status_enabled.where("end_date IS NULL OR end_date > ?", now) }
   scope :recording, ->(now = Time.now) { active.where("start_date > ?", now) }
@@ -35,7 +36,6 @@ def self.config
         end_date: s.end_date,
         record_days: s.record_days,
         record_hours: s.record_hours,
-        job_id: "#{s.podcast_id}/#{s.id}/:date/:hour",
         callback: PorterUtils.callback_sqs
       }
     end
diff --git a/app/models/stream_resource.rb b/app/models/stream_resource.rb
new file mode 100644
index 000000000..dcb65568d
--- /dev/null
+++ b/app/models/stream_resource.rb
@@ -0,0 +1,71 @@
+class StreamResource < ApplicationRecord
+  BUFFER_SECONDS = 10
+
+  enum :status, %w[started created processing complete error retrying cancelled invalid].to_enum_h, prefix: true
+
+  belongs_to :stream_recording, -> { with_deleted }, touch: true, optional: true
+  has_one :podcast, through: :stream_recording
+  has_one :task, -> { order(id: :desc) }, as: :owner
+  has_many :tasks, as: :owner
+
+  validates :start_at, presence: true
+  validates :end_at, presence: true, comparison: {greater_than: :start_at}
+  validates :actual_start_at, presence: true
+  validates :actual_end_at, presence: true, comparison: {greater_than: :actual_start_at}
+  validates :original_url, presence: true
+
+  after_initialize :set_defaults
+  before_validation :set_defaults
+
+  acts_as_paranoid
+
+  # find/build for a specific oxbow job_id
+  # <podcast_id>/<stream_recording_id>/<start_time>/<end_time>/<guid>.mp3
+  def self.decode(str)
+    parts = str.split("/")
+    podcast_id = parts[0].to_i
+    recording_id = parts[1].to_i
+    start_at = safe_parse_time(parts[2])
+    end_at = safe_parse_time(parts[3])
+    return unless podcast_id > 0 && recording_id > 0 && start_at && end_at
+
+    rec = StreamRecording.find_by_id(recording_id)
+    return unless rec
+
+    res = rec.stream_resources.find_by(start_at: start_at, end_at: end_at)
+    return res if res
+
+    rec.stream_resources.build(start_at: start_at, end_at: end_at)
+  end
+
+  # NOTE: called twice, because podcast won't be there on initialize
+  def set_defaults
+    set_default(:status, "created")
+    set_default(:guid, SecureRandom.uuid)
+    set_default(:url, published_url)
+  end
+
+  def file_name
+    File.basename(URI.parse(original_url).path) if original_url.present?
+  end
+
+  def published_path
+    "#{podcast.path}/#{stream_resource_path}" if podcast
+  end
+
+  def published_url
+    "#{podcast.base_published_url}/#{stream_resource_path}" if podcast
+  end
+
+  private
+
+  def stream_resource_path
+    "streams/#{guid}/#{file_name}"
+  end
+
+  def safe_parse_time(str)
+    str&.to_time
+  rescue
+    nil
+  end
+end
diff --git a/db/migrate/20251218154550_create_stream_resources.rb b/db/migrate/20251218154550_create_stream_resources.rb
new file mode 100644
index 000000000..7f3c9c1c9
--- /dev/null
+++ b/db/migrate/20251218154550_create_stream_resources.rb
@@ -0,0 +1,30 @@
+class CreateStreamResources < ActiveRecord::Migration[7.2]
+  def change
+    create_table :stream_resources do |t|
+      t.references :stream_recording, index: true, foreign_key: true
+
+      # timeframe we were trying to capture, vs actually recorded
+      t.timestamp :start_at, index: true
+      t.timestamp :end_at, index: true
+      t.timestamp :actual_start_at
+      t.timestamp :actual_end_at
+
+      # file locations
+      t.string :guid
+      t.string :url
+      t.string :original_url
+
+      # metadata
+      t.string :status
+      t.string :mime_type
+      t.integer :file_size
+      t.integer :bit_rate
+      t.decimal :sample_rate
+      t.integer :channels
+      t.decimal :duration
+
+      t.timestamps
+      t.timestamp :deleted_at
+    end
+  end
+end
diff --git a/db/schema.rb b/db/schema.rb
index afc9565ec..dc6cedf44 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[7.2].define(version: 2025_11_19_222707) do
+ActiveRecord::Schema[7.2].define(version: 2025_12_18_154550) do
   # These are extensions that must be enabled in order to support this database
   enable_extension "plpgsql"
   enable_extension "uuid-ossp"
@@ -476,6 +476,30 @@
     t.index ["podcast_id"], name: "index_stream_recordings_on_podcast_id"
   end
 
+  create_table "stream_resources", force: :cascade do |t|
+    t.bigint "stream_recording_id"
+    t.datetime "start_at", precision: nil
+    t.datetime "end_at", precision: nil
+    t.datetime "actual_start_at", precision: nil
+    t.datetime "actual_end_at", precision: nil
+    t.string "guid"
+    t.string "url"
+    t.string "original_url"
+    t.string "status"
+    t.string "mime_type"
+    t.integer "file_size"
+    t.integer "bit_rate"
+    t.decimal "sample_rate"
+    t.integer "channels"
+    t.decimal "duration"
+    t.datetime "created_at", null: false
+    t.datetime "updated_at", null: false
+    t.datetime "deleted_at", precision: nil
+    t.index ["end_at"], name: "index_stream_resources_on_end_at"
+    t.index ["start_at"], name: "index_stream_resources_on_start_at"
+    t.index ["stream_recording_id"], name: "index_stream_resources_on_stream_recording_id"
+  end
+
   create_table "subscribe_links", force: :cascade do |t|
     t.datetime "created_at", null: false
     t.datetime "updated_at", null: false
@@ -543,4 +567,5 @@
   add_foreign_key "publishing_pipeline_states", "publishing_queue_items"
   add_foreign_key "publishing_queue_items", "podcasts"
   add_foreign_key "stream_recordings", "podcasts"
+  add_foreign_key "stream_resources", "stream_recordings"
 end
diff --git a/test/factories/oxbow_callback_factory.rb b/test/factories/oxbow_callback_factory.rb
new file mode 100644
index 000000000..0c51f0b54
--- /dev/null
+++ b/test/factories/oxbow_callback_factory.rb
@@ -0,0 +1,79 @@
+FactoryBot.define do
+  factory :oxbow_job_id, class: String do
+    Id { "1234/5678/2025-12-17T15:00Z/2025-12-17T16:00Z/27a8112d-b582-4d23-8d73-257e543d64a4.mp3" }
+    initialize_with { attributes.with_indifferent_access }
+  end
+
+  factory :oxbow_job_received, class: Hash do
+    Time { "2025-12-17T14:51:16.490Z" }
+    Timestamp { 1765983076.49 }
+    JobReceived do
+      Job { build(:oxbow_job_id) }
+      Execution { {Id: "arn:aws:states:the-execution-arn"} }
+      State "RECEIVED"
+    end
+    initialize_with { attributes.with_indifferent_access }
+  end
+
+  factory :oxbow_ffmpeg_destination, class: Hash do
+    Mode { "AWS/S3" }
+    BucketName { "prx-feed-testing" }
+    ObjectKey { "1234/5678/2025-12-17T15:00Z/2025-12-17T16:00Z/27a8112d-b582-4d23-8d73-257e543d64a4.mp3" }
+    initialize_with { attributes.with_indifferent_access }
+  end
+
+  factory :oxbow_ffmpeg_task, class: Hash do
+    Type { "FFmpeg" }
+    FFmpeg do
+      Inputs { "-t 4350 -i \"http://some.stream/url/stream_name.mp3\"" }
+      GlobalOptions { "" }
+      InputFileOptions { "" }
+      OutputFileOptions { "" }
+      Outputs { [{Format: "mp3", Destination: build(:oxbow_ffmpeg_destination)}] }
+    end
+    initialize_with { attributes.with_indifferent_access }
+  end
+
+  factory :oxbow_ffmpeg_result, class: Hash do
+    Task { "FFmpeg" }
+    Time { "2025-12-17T16:02:23.826Z" }
+    Timestamp { 1765987343.826 }
+    FFmpeg do
+      Outputs do
+        [
+          Mode: "AWS/S3",
+          BucketName: "prx-feed-testing",
+          ObjectKey: "1234/5678/2025-12-17T15:00Z/2025-12-17T16:00Z/27a8112d-b582-4d23-8d73-257e543d64a4.mp3",
+          Duration: 4350000,
+          Size: 12345678
+        ]
+      end
+    end
+    initialize_with { attributes.with_indifferent_access }
+  end
+
+  factory :oxbow_task_result, class: Hash do
+    Time { "2025-12-17T16:02:23.210Z" }
+    Timestamp { 1765987343.21 }
+    Task { build(:oxbow_ffmpeg_task) }
+    TaskResult do
+      Job { build(:oxbow_job_id) }
+      Execution { {Id: "arn:aws:states:the-execution-arn"} }
+      Result build(:oxbow_ffmpeg_result)
+    end
+    initialize_with { attributes.with_indifferent_access }
+  end
+
+  factory :oxbow_job_results, class: Hash do
+    Time { "2025-12-17T16:02:23.826Z" }
+    Timestamp { 1765987343.826 }
+    JobResult do
+      Job { build(:oxbow_job_id) }
+      Execution { {Id: "arn:aws:states:the-execution-arn"} }
+      State { "DONE" }
+      FailedTasks { [] }
+      TaskResults { [build(:oxbow_ffmpeg_result)] }
+    end
+    initialize_with { attributes.with_indifferent_access }
+  end
+end
diff --git a/test/factories/stream_resource_factory.rb b/test/factories/stream_resource_factory.rb
new file mode 100644
index 000000000..64afb0759
--- /dev/null
+++ b/test/factories/stream_resource_factory.rb
@@ -0,0 +1,18 @@
+FactoryBot.define do
+  factory :stream_resource do
+    start_at { "2025-12-18T13:00:00Z" }
+    end_at { "2025-12-18T14:00:00Z" }
+    actual_start_at { "2025-12-18T12:59:50Z" }
+    actual_end_at { "2025-12-18T14:00:10Z" }
+
+    original_url { "s3://prx-testing/test/audio.mp3" }
+
+    status { "complete" }
+    mime_type { "audio/mpeg" }
+    file_size { 774059 }
+    bit_rate { 128 }
+    sample_rate { 44100 }
+    channels { 2 }
+    duration { 3620.0 }
+  end
+end
diff --git a/test/models/stream_resource_test.rb b/test/models/stream_resource_test.rb
new file mode 100644
index 000000000..19bdb14a4
--- /dev/null
+++ b/test/models/stream_resource_test.rb
@@ -0,0 +1,92 @@
+require "test_helper"
+
+describe StreamResource do
+  let(:podcast) { build_stubbed(:podcast) }
+  let(:resource) { build_stubbed(:stream_resource, podcast: podcast) }
+
+  describe ".decode" do
+    let(:rec) { create(:stream_recording) }
+    let(:rec2) { create(:stream_recording) }
+
+    it "finds or builds stream resources" do
+      id = "1234/#{rec.id}/2025-12-17T15:00Z/2025-12-17T16:00Z/some-guid.mp3"
+      res = StreamResource.decode(id)
+
+      assert res.new_record?
+      assert_equal rec.id, res.stream_recording_id
+      assert_equal "2025-12-17T15:00Z".to_time, res.start_at
+      assert_equal "2025-12-17T16:00Z".to_time, res.end_at
+
+      # skip validations - we're just testing finding my stream+dates
+      res.save(validate: false)
+      assert res.persisted?
+
+      # finding again returns the same resource
+      assert_equal res, StreamResource.decode(id)
+      assert_equal res, StreamResource.decode(id.sub("some-guid.mp3", "other-guid.mp3"))
+
+      # but different recordings or dates are not new resources
+      id2 = id.sub("/#{rec.id}/", "/#{rec2.id}/")
+      assert StreamResource.decode(id2).new_record?
+      assert StreamResource.decode(id.sub("15:00Z", "15:30Z")).new_record?
+      assert StreamResource.decode(id.sub("16:00Z", "16:05Z")).new_record?
+
+      # bad data is nil
+      assert_nil StreamResource.decode(id.sub("1234/", "abcd/"))
+      assert_nil StreamResource.decode(id.sub("/#{rec.id}/", "/abcd/"))
+      assert_nil StreamResource.decode(id.sub("17T15:00Z", ""))
+      assert_nil StreamResource.decode(id.sub("17T16:00Z", ""))
+      assert_nil StreamResource.decode("1234/5678/something")
+      assert_nil StreamResource.decode("whatev")
+    end
+  end
+
+  describe "#set_defaults" do
+    it "sets unchanged defaults" do
+      res = StreamResource.new(podcast: podcast)
+      assert_equal "created", res.status
+      assert res.guid.present?
+      assert res.url.present?
+      refute res.changed?
+    end
+  end
+
+  describe "#file_name" do
+    it "parses the original url" do
+      assert_equal "audio.mp3", resource.file_name
+
+      resource.original_url = "http://some.where/the/file.name.here#and?other&stuff=1"
+      assert_equal "file.name.here", resource.file_name
+
+      resource.original_url = ""
+      assert_nil resource.file_name
+    end
+  end
+
+  describe "#published_path" do
+    it "includes the podcast prefix" do
+      assert_equal "#{podcast.id}/streams/#{resource.guid}/audio.mp3", resource.published_path
+    end
+  end
+
+  describe "#published_url" do
+    it "includes the podcast http url" do
+      assert_equal "https://f.prxu.org/#{resource.published_path}", resource.published_url
+    end
+
+    it "sets the url field" do
+      # NOTE: podcast is nil after_initialize
+      assert_nil resource.url
+
+      # but will be there before validation
+      assert resource.valid?
+      refute_nil resource.url
+      assert_equal resource.published_url, resource.url
+
+      resource.guid = "some-other-guid"
+      resource.original_url = "http://some/other.filename"
+      assert resource.valid?
+      refute_equal resource.published_url, resource.url
+    end
+  end
+end