Home › Features › Projects & work

Area 01

Projects & work

A project overview, collaborative knowledge and a task board that stays focused.

Projects The project wiki Tasks Datasets

Ask for it

Turn meeting notes into connected tasks.

Your assistant reads the note, creates tasks in the right project and links each one back to the decision it came from.

Projects

The whole project on one page.

See the team, capacity, time, deadlines, tasks, customer context and project knowledge in one place.

Compare planned and tracked time before the project runs over budget.

A project detail page: a team and capacity table, planned against tracked time, and wiki and task widgets.

The project wiki

Where the work gets written down.

Every project has a wiki for meeting notes, decisions and product briefs — written together, connected to tasks and shared only where you choose.

A wiki page in reading view: page tree with nested subpages, breadcrumbs, an embedded interactive flow chart, and linked task pills.
The wiki editor on a meeting note: attendee chips, the Markdown text and an insert menu offering a task, a page, an image and a flow chart.
Everyone edits the same page at once Diagrams and tables directly in the page A linked task, straight from the note
  • Collaborative project knowledgeEdit pages together and keep meeting notes, decisions and product briefs in one place.
  • Turn decisions into workCreate linked tasks directly from meeting notes.
  • Share only what a client needsGive guests access to selected pages without exposing the rest of the project.
  • Discuss work in contextComment on exact passages and keep the conversation with the source material.
  • Use rich content without lock-inAdd diagrams, tables and files, then import or export in Markdown.

Tasks

A board that stays small.

Types, six statuses — "Won't do" is an answer too — one assignee and one level of subtasks. Small enough that the board stays a view of the work.

The Tasks board: status columns with typed task cards, assignees and due dates.
A task opened read-first: the description up front, a P0–P3 priority control and quick status and owner selects in the sidebar, and the Weekly sync meeting note it came from linked below.

Compact mode fits the whole sprint on one screen.

The same board in compact mode: one line per card with the key, the title, a due date and the owner's avatar, and every status column visible at once.
  • A focused task boardTrack status, priority, ownership and deadlines without turning the board into a second job.
  • Plan in sprintsKeep current work together and carry unfinished items forward.
  • Connect code and workLink GitHub activity to the tasks it belongs to.
  • Keep the contextFiles, comments and linked wiki pages stay with the task.

Datasets

A row a day, next to the roadmap.

Some of the numbers a team lives by are written by machines — the app's store figures every morning, a monitor checking the site every five minutes, an assistant reporting what it read overnight. Each gets a dataset under its project, appends one row at a time, and the wiki shows the history as a live table or chart.

A dataset of the mobile app's daily store metrics: one profile card per column with a histogram or the most common values, the sortable table of rows beneath, and the ingest token with its ready-made curl snippet beside it.
A wiki page with the same dataset embedded as a live chart of daily actives and a filtered table of Android stability beneath a paragraph of commentary. 
POST /datasets/12/rows
Authorization: Bearer opds_…
Idempotency-Key: ios-2026-08-22
Content-Type: application/json

{ "timestamp": "2026-08-22T06:30:00+02:00", "platform": "ios",
  "installs": 113, "dau": 3359, "crash_free_pct": 99.8, "rating": 4.4 }

→ 201  { "inserted": 1, "duplicates": 0, "ids": [4471] }
  • Append, never re-uploadEvery record is one row added with one request — there is no file to fetch, edit and put back, so two writers can never overwrite each other. Send an Idempotency-Key and a retried request is stored once.
  • A token that can do one thingEach dataset has its own ingest token, and all it can do is append rows to that dataset. Rotate or revoke it from the dataset page, which also hands you the curl, RouterOS and Python snippets ready to run.
  • Free-form, or declaredRows are JSON. Declare the columns and their types and a malformed record is refused with the key named; leave them undeclared and the dataset takes what comes — an uptime monitor's status and response time, say.
  • Profiled, charted, in the wikiThe dataset opens on a column profile — type, count, a histogram or the most common values — above a table that sorts and filters on any column and a time range, a chart of the numeric columns, and JSONL, CSV or JSON exports. Save a sort, filters and a chart as a named view, then type /dataset in a page to embed it: it reads the current rows every time the page opens, and stays as private as the page.
  • Written by an assistant, tooA scheduled assistant that scans the news, social media and research papers each morning appends what it found — source, title, relevance — and the team reads the table in the wiki. The same goes for a CI job, a router or a sensor: anything that can make one request can keep a dataset.
  • Bounded on purposeA dataset keeps up to 100 000 rows and ages old ones out on the retention you choose. A secret download link for a dashboard outside Opelli is off until you switch it on, and says plainly what holding it means.

Ask for it

From the day's rows to a chart on the page.

Your assistant finds the dataset, reads the month out of it, saves what it read as a named view and puts that view on the wiki page — as a live chart that keeps growing with each morning's row.

See it on your own operations.

Opelli runs in the cloud on an opelli.dev address or your own domain. Leave your email and we will show you how it fits the way your team works.

NextPlanning & people →