# Can Codex run a feature vote on Threads and count the replies with a script?

> Source: https://tareno.co/agents/codex/threads
> Language: en
> Last reviewed: 2026-10-24
> Use Codex with Tareno to post a feature vote on Threads, read the replies and count them with a small script that counts each person once, with approval in Tareno.

## Connection facts

| Field | Value |
| --- | --- |
| Product surface | Codex (codex) |
| Available connection | Published plugin |
| Connection evidence | documented |
| Methods | plugin, mcp |
| Authentication | oauth |
| Setup | [Setup guide](https://tareno.co/docs/mcp/openai) |
| Install or connect | [Open setup](https://chatgpt.com/plugins/plugin_asdk_app_6a54fa8fd95c81918e64891984183432) |
| Evidence observed | 2026-10-27 |
| Facts reviewed | 2026-10-27 |

Publication status and a working connection are separate facts. Evidence for one product surface does not establish installation or compatibility in another surface.

Agent: Codex · Network: Threads · Hub: [Codex and Tareno](https://tareno.co/agents/codex)

## Short answer

**Can Codex post a feature vote on Threads and count the replies?**

Yes. Codex writes the vote post within 500 characters, saves it as a draft through Tareno, and later reads the replies and tallies them with a small script that counts each person once. Threads has no poll feature in Tareno, so votes are replies. Scheduling stays pending until you approve it in Tareno.

Asking users what to build next is a good habit, and Threads is a natural place to ask because people answer there. The trouble is counting. Without a poll feature, votes arrive as replies in many forms: a letter, a feature name, a sentence of explanation, a second vote from the same person. Codex is good at the small, boring script that turns that mess into numbers, and at saying where the script is unsure. This page follows the maintainer of an open-source note-taking app through a vote on three features, with the counting rules that keep the result honest and the caution that a reply count is a signal, not a mandate.

## Example: a note-taking app asks which feature to build next

Imagine a maintainer with three candidates: offline sync, tags and a dark theme. Codex drafts a post of 280 characters that asks users to reply with A, B or C, and saves it with `create_drafts` for the project's Threads profile. Two days after it is live she asks Codex to read the replies with `list_post_comments`. It writes a script that takes the first valid vote per person, reads a letter, a feature name or a clear sentence, and sets unclear replies aside. The tally shows 41 for sync, 23 for tags and 12 for the theme, with nine replies it could not classify, which she reads herself.

## Step by step

1. **Define the options and the rules.** Choose three options at most and decide how people vote, for example by replying with a letter. Fewer options and one rule produce replies a script can read.
2. **Draft the vote post.** Codex writes a post within 500 characters that names the options and the voting rule. Meta counts emojis by bytes, so keep decoration minimal.
3. **Save and approve.** Codex confirms the profile with `list_accounts` and saves the draft through `create_drafts`. Give a day, a time and an IANA time zone, and approve the pending request in the Tareno dashboard.
4. **Read the replies.** After a day or two, Codex reads the replies with `list_post_comments`. Threads is a network where Tareno reads replies, which is why this poll-by-replies design works here.
5. **Write the tally script.** Ask Codex for a small script that counts one vote per person, recognizes a letter or a feature name, sets unclear replies aside and prints the totals and the unclassified list.
6. **Read what the script could not.** Review the unclassified replies yourself. Some will be good ideas you did not offer, and those are worth more than the count.

## Prompts to copy

### Draft the vote

```text
Write a Threads post within 280 characters that asks users of my note-taking app to vote for the next feature by replying with A, B or C: A offline sync, B tags, C dark theme. State that only the first vote per person counts, and add no emojis.
```

### Save through Tareno

```text
Use Tareno to confirm my Threads profile [project handle] and save the approved vote post as a draft. Return the draft ID and the text, and do not request a time until I give you the day and the time zone.
```

### Tally the replies

```text
Use Tareno to read the replies to the vote post and write a small script that counts one vote per person, recognizes A, B, C or the feature name, and sets unclear replies aside. Run it, print the totals and list the unclassified replies with a one-line reason for each.
```

### Report honestly

```text
Write a short follow-up post within 500 characters that reports the aggregate result, says how many replies were counted and how many could not be classified, and states that the vote guides the decision rather than making it. Name no one who voted.
```

## Threads in Tareno

- Tareno has no poll feature for Threads, so a vote is a post that asks for replies, and the counting is a script over the replies Tareno reads.
- Meta's Threads documentation limits text posts to 500 characters, so a vote with three options and a rule fits easily if the options are named briefly.
- Meta's documentation limits a profile to 250 API-published posts within 24 hours, which a vote never approaches but a scripted job should still respect.
- Links get an automatic preview on Threads in Tareno's configuration, so the roadmap can be linked from the vote post.

Official rules ([Meta: Threads API, posts](https://developers.facebook.com/docs/threads/posts), checked 2026-10-24):

- A profile is limited to 250 API-published posts within 24 hours; a carousel counts as one post.
- Text posts are limited to 500 characters, and emojis count by their UTF-8 bytes, not as single characters.
- Carousels need 2 to 20 items. Single images may be JPEG or PNG up to 8 MB, and videos MOV or MP4 up to 5 minutes and 1 GB.

## Mistakes to avoid

- Counting every reply as a vote. Count one per person, or a few enthusiasts decide the result.
- Offering too many options. Three choices keep replies readable.
- Treating the result as a mandate. A reply count reflects the people who follow you, not all users.
- Naming voters in the report. Publish aggregates only.

## What does not work

- Codex cannot guarantee that each voter is a distinct person, because replies identify accounts and not people.
- It cannot post replies on Threads, since Tareno reads replies but has no tool that publishes one.
- It cannot publish the vote on its own, since a time request becomes a post only after you approve it in the Tareno dashboard.

## Frequently asked questions

### Does Threads have a poll feature in Tareno?

No. Tareno has no poll tool for Threads, so the vote is a post that asks for replies. That works because Tareno reads replies, and Codex can count them with a small script.

### How does the script avoid double counting?

It keeps the first valid vote per account and ignores later ones. That is a rule you can state in the post. Accounts are not people, so treat the result as a signal and not as an election.

### What about replies that do not follow the format?

The script sets them aside and lists them with a reason. Read them yourself, because some are feature ideas you did not offer, and those can matter more than the votes.

### Should I publish the numbers?

Publish the aggregate and the number of replies you could not classify, and say that the result guides the decision. Do not name voters, and do not present a small count as the will of all users.

### Can the script run on a schedule?

It can, but an unattended run should only read replies and print totals. Keep posting and scheduling out of it, so the follow-up always passes your approval in the dashboard.

### Will the vote post go out without my approval?

No. A scheduling request is a pending action in Tareno, and Threads receives the post only after you approve it in the dashboard. `get_action_status` shows what happened afterwards.

## Tareno tools used

- `list_post_comments`
- `list_accounts`
- `create_drafts`
- `get_action_status`

Approval model: Tareno saves drafts directly. Scheduling, publishing, edits to scheduled posts and deletions become pending actions that a person approves in the Tareno dashboard; no agent can approve them.

## Related

- [Codex × Tareno overview](https://tareno.co/agents/codex)
- [Threads with Tareno](https://tareno.co/channels/threads)
- [Post scheduling](https://tareno.co/features/post-scheduling)
- [Approval workflows](https://tareno.co/features/approval-workflows)
- [Codex and ChatGPT setup](https://tareno.co/docs/mcp/openai)