AlloFlow Packets All Packets pages
This page uses an AI voice, on this device. No account. Your passage is not uploaded. See the details Hide the details
  • This is one of the two Packets pages that use AI. The other is Make recordings with a computer voice. The other pages do not, and nothing on them changes.
  • The voice is Kokoro, a computer voice made with AI. It runs inside this browser tab. The passage you type is turned into sound on this device and is not sent anywhere.
  • The words are yours. This page does not write, shorten or change the passage.
  • Nothing is downloaded until you press Get the voice. Then the page downloads the voice, about 88 MB, one time, from Hugging Face (huggingface.co). Hugging Face can see that your browser asked for the voice. It does not receive your passage.
  • The program that runs the voice comes from this site (or AlloFlow's CDN at alloflow-cdn.pages.dev when hosted on a mirror without the engine files), and the page checks it before running it.
  • If you press Record on a sentence, this page asks for this device's microphone to record your voice inside this tab. Your recording stays on this device, is not stored and is not uploaded.
  • The passage and the recordings stay on this page only while it is open. They are not stored and not uploaded. They go only into the file you save.
  • This page loads sixteen files from this site: the page itself, five stylesheets (packets.css, packets-make.css, packets-voice.css, packets-read-along.css, packets-read-along-print.css), ten scripts (packets-page.js, word_sounds_take_home_loader.js, packet_read_along_builder.js, syllable_table_en.js, packets_voice_engine.js, packets_verify_core.js, packets_reopen.js, packets_handoff.js, packets_zip.js, packets-read-along.js) and the tab icon. When you record or use the computer voice, it loads packets_audio_kit.js and lame.min.js from this site to make recordings smaller.
  • There are no cookies and nothing to sign in to.

The file you save can:

  • Show the passage a page at a time, and play your recordings of it, one sentence at a time, with the sentence highlighted.
  • Play one sentence again when the student presses it, and play more slowly.
  • Show longer words in syllables when the student presses Syllables. The breaks come from a word list on this site, looked up here when you make the file. No AI is used for them.
  • Offer the Reading tools every Packets file has: text size, colors, spacing, letters, a reading focus line and a magnifier.
  • Keep the student's place on that device.

The file you save cannot:

  • Run AI. It makes no speech: it only plays the recordings that were made here.
  • Connect to the internet. The file tells the browser to refuse every connection, and the browser enforces it.
  • Listen to the student, unless you turn on optional student voice practice in step 4 (which stays in the student's browser tab only and is never saved or sent). It never uses the camera.

The rule this page gives your browser. The browser refuses anything the rule does not allow: It is the Content-Security-Policy line near the top of this page's source.

Make a read-along

Type or paste a passage. A computer voice records it one sentence at a time, and you save one file that reads it to a student with each sentence highlighted as it is read. Listening practice: nothing is scored.

Step 1

Your passage

The student sees it at the top of the file. Left empty, it is "Read along".
Up to 6,000 letters and 120 sentences. Each sentence can have its own recording, or you can save a reading file with Syllables and Reading tools without recordings. To choose the breaks yourself, put each sentence on its own line. An empty line starts a new paragraph.

Show the whole passage on one scrollable page, split automatically into pages, or choose your own page breaks (put --- on its own line, or use Split page in step 3). Start a line with # for a heading.
Right-to-left languages set the file direction automatically. Record non-English passages with your microphone in step 3.

Step 2

The voice (optional)

Use a computer voice here, record your own voice in step 3 with this device's microphone, or skip recordings to make a reading file with Reading tools and Syllables.

The voice reads English. A recording keeps the voice it was made with.

Downloaded one time from Hugging Face, then kept by this browser. On a slow connection this takes a few minutes.

Step 3

Make the recordings (optional)

One recording for each sentence, using your microphone or the computer voice. Listen to each one: a computer voice can say a word wrong. Make it again, or change the sentence in step 1.

Type a passage in step 1 first.

    Step 4

    Save the file

    Type a passage in step 1 to build the file.

    Put student code names in the file

    Type 1 code name to put it into the file above, or 2 to 40 (one per line) to save one file per student in a .zip. In Google Classroom, where one assignment gives every student the same file, leave this empty and each student types a code name at the start.

    Use code names or first names only, never full names. At most 40 at a time.

    Sharing by web link instead? Add #learner=CodeName to the end of the link, for example #learner=BlueHeron7. It stays on the student's device.