UneedLists — new product launches

Quill

Lucas Ferreira — Sat, 15 Aug 2026 07:00:00 GMT

A writing environment where the outline is a real object, not a heading hierarchy

Quill is built for long-form work — books, documentation, research, long essays — where structure matters more than word count. The outline is a first-class, manipulable object, so restructuring 40,000 words is a drag rather than an afternoon of cutting and pasting.

Features

Outline as an object — Drag to restructure; cross-references follow automatically.
Section status — Outline, draft, revised, done — see the state of a manuscript at a glance.
Fold to any depth — Read 60,000 words as a one-page outline and descend selectively.
Durable cross-references — References point at sections, so rewriting a heading does not break them.
Plain files on disk — Your manuscript is a folder of Markdown. Quill is a lens, not a silo.
Real export — Markdown, LaTeX, DOCX and EPUB, all of which actually compile.

Long-form writing is a structure problem Below about 3,000 words, structure takes care of itself. Above 20,000, structure is the entire job, and every general-purpose editor treats it as a side effect of your heading levels. Quill makes the outline the primary object and the prose secondary. What that means in practice - Restructure by dragging. Move a section and everything under it moves, with cross-references updating automatically. - Write out of order. Sections carry status — outline, draft, revised, done — so a half-finished manuscript is legible at a glance instead of being a wall of un…

Read more on UneedLists

Halden

Tobias Lind — Fri, 14 Aug 2026 07:00:00 GMT

A self-hosted CRM for consultancies that will not put client data in the cloud

Halden is a small, self-hosted CRM built for agencies and consultancies with confidentiality obligations. One binary, one Postgres database, no telemetry, and no AI features reading your client correspondence. Small enough that one person can audit the entire codebase.

Features

One binary, one database — Postgres and a single process. No Redis, no queue, no orchestrator.
Zero telemetry — No outbound connections you did not configure, in a codebase small enough to verify.
Retainer-aware — Recurring engagements and time tracking that feed invoicing, not just deal stages.
No AI on your correspondence — Nothing reads client email. Stated as a product commitment, not a current limitation.
Skip-version upgrades — Upgrade from a year-old install in one command.
Auditable — 14,000 lines. You can read the whole thing.

Built for the confidentiality constraint Some businesses cannot put client data in a multi-tenant SaaS. Law-adjacent consultancies, agencies under NDA, anyone with a client whose procurement team asks where the data physically lives. The usual options are an enterprise CRM that costs more than the team, or a spreadsheet. Halden is the third option. Deliberately small - One binary. Drop it on a VPS, point it at Postgres, done. No Kubernetes, no message queue, no Redis. - No telemetry. Halden makes zero outbound connections you did not configure. This is verifiable — the codebase is 14,000 lines…

Read more on UneedLists

Driftwood

Kenji Watanabe — Thu, 13 Aug 2026 07:00:00 GMT

A local-first sync engine you can drop into an app you already shipped

Driftwood handles offline queues, conflict resolution and live schema migration so your app works on a plane and reconciles cleanly when it lands. Open source, framework-agnostic, and designed to be added to an existing data model rather than requiring a rewrite around it.

Features

Retrofit, do not rewrite — Attaches to an existing schema instead of demanding you adopt a new data model.
Per-field conflict strategy — LWW, CRDT merge or your own callback — chosen per field, not per database.
Live schema migration — Forward and backward transforms so old and new app versions coexist safely.
Partial sync — Clients hold the subset they need. Datasets are not capped by device storage.
Durable offline queue — Ordered, idempotent replay that survives crashes and force-quits.
Framework agnostic — SQLite, IndexedDB and Postgres adapters. No opinion about your UI layer.

Sync is the hard part of local-first Local-first software gets described as a UX philosophy. Operationally it is a distributed systems problem: two replicas, unreliable connectivity, concurrent edits, and schema versions that are live at the same time because users do not update apps on your schedule. Driftwood is the part nobody enjoys writing. What it handles - Offline queues with durable, ordered, idempotent replay. - Conflict resolution with per-field strategies — last-writer-wins where it is fine, CRDT merge where it is not, and an explicit callback where only your domain knows the answer…

Read more on UneedLists

Shipyard

Priya Raman — Wed, 12 Aug 2026 07:00:00 GMT

Deploy previews for data pipelines, so reviewers see the diff in the data

Shipyard gives every pull request an isolated, disposable warehouse built from a sample of production. Reviewers see what a SQL change actually does to the numbers — row counts, distributions, breaking column changes — instead of reading the query and hoping.

Features

Per-PR warehouses — Isolated, disposable, built from a stratified production sample.
Data diffs in review — Row counts, distribution shifts and schema changes posted to the pull request.
Downstream impact — Names every model whose output changed, not just the one you edited.
Cheap by construction — Sampling keeps a preview environment at cents rather than dollars.
dbt and SQLMesh native — Existing projects work without modification.
Per-repo pricing — Add reviewers without adding cost.

Data teams review SQL, not results Application teams get a preview environment on every pull request. Data teams get a shared staging warehouse, a queue for it, and a review process where someone reads 200 lines of SQL and approves based on vibes. Shipyard closes that gap. What happens on a pull request 1. Shipyard builds an isolated warehouse from a stratified sample of production — small enough to be cheap, representative enough to be meaningful. 2. Your models run against it. 3. The PR gets a comment showing the data diff: row count deltas, distribution shifts, new and dropped columns, and…

Read more on UneedLists

Tempo

Mira Okonkwo — Tue, 11 Aug 2026 07:01:00 GMT

Your calendar, rewritten every morning so you make fewer decisions

Tempo reads your calendar, tasks and energy patterns, then rebuilds your day into focused blocks before you wake up. It is not another planner you have to maintain — it maintains itself, and it defends your deep-work time against the meetings that would otherwise erode it.

Features

Overnight rescheduling — Your day is rebuilt at 4am against fresh calendar state — no morning triage.
Meeting defence — Auto-decline or counter-propose invitations that would fragment protected focus blocks.
Energy-aware placement — Learns when you actually complete demanding work and places it there.
Honest capacity — Refuses to overbook. If the day is full, Tempo tells you what will not fit.
Two-way calendar sync — Conflict-aware sync with Google, Outlook and CalDAV. Your manual edits always win.
Keyboard-first — Every action reachable without a mouse. Opens in under 200ms.

The problem with planners Every planning tool asks you to do the planning. That is the part that is hard, and it is the part that silently stops happening around week three. Tempo inverts it. Overnight, it looks at what is on your calendar, what is on your task list, when you have historically done your best work, and it produces a schedule. You wake up to a day that is already arranged. How it works Tempo runs three passes over your day: 1. Protect — deep-work blocks are placed first, in the windows where your completed-task history says you actually focus. 2. Fit — tasks are slotted against…

Read more on UneedLists

Anchor

Sana Iqbal — Mon, 10 Aug 2026 07:00:00 GMT

An AI research assistant that shows its sources or says it does not know

Anchor answers research questions against your own document corpus and cites the exact passage behind every claim. When the corpus does not contain an answer it says so instead of generating a plausible one — which is the only behaviour that makes an AI research tool usable for work that matters.

Features

Passage-level citations — Every claim links to the exact paragraph behind it, not just the document.
Honest misses — Returns "I don't have that" rather than falling back to model knowledge.
Self-hosted embeddings — A corpus under NDA never leaves your network.
Confidence floor — A tunable threshold below which Anchor declines to answer at all.
No persona — A research interface, not a chatbot. Query in, cited answer out.
Corpus diffing — Shows which answers changed when you add or remove documents.

The problem is not fluency Language models are already fluent enough. The failure mode that matters for research is that a confident wrong answer costs more than no answer at all, and you cannot tell them apart by reading. Anchor is built around making that distinction visible. How it works Every claim in an Anchor answer is anchored to a span in a source document — not a document-level citation, a passage-level one. Click any sentence and you land on the exact paragraph it came from. When retrieval returns nothing above the confidence floor, Anchor returns "I don't have that" and shows you wh…

Read more on UneedLists

Overcast

Daniel Reyes — Sun, 09 Aug 2026 07:02:00 GMT

Observability a two-person team can afford to run and actually understand

Overcast is metrics, logs and traces for small teams, priced and designed for engineers who are also the on-call rotation. One binary, one Postgres database, and alert rules you can read out loud. No sampling surprises, no per-host pricing, no bill that outgrows your infrastructure.

Features

One binary — Metrics, logs and traces from a single process backed by Postgres and S3-compatible storage.
Readable alert rules — Rules are written the way you would say them, and they diff cleanly in code review.
No silent sampling — Overcast never quietly drops data. Budget overruns are loud and configurable.
OpenTelemetry native — Standard OTLP ingest. Your instrumentation is portable on day one and on the day you leave.
Correlated by default — Jump from an alert to the trace to the exact log line without changing tools.
Self-host the real build — The open-source build is the production build. No feature-gated community edition.

Built for the team that is also the on-call rotation Observability vendors design for a buyer who has a platform team. If that is not you, you get one of two outcomes: a free tier that expires exactly when it becomes useful, or a bill that dwarfs the infrastructure it is watching. Overcast is built for the middle. What you get - Metrics, logs and traces in one place, correlated by trace ID, with no separate products to wire together. - Alert rules in plain language. is the actual rule syntax. - Predictable pricing. Priced on ingested gigabytes, not hosts, containers, seats or custom metrics. -…

Read more on UneedLists

Northlight

Sana Iqbal — Thu, 06 Aug 2026 07:00:00 GMT

Design tokens that stay honest between Figma and production code

Northlight makes your codebase the source of truth for design tokens and generates the Figma side from it. Colour, type, spacing and motion stay identical across both, and drift becomes a failing CI check instead of a discovery three sprints later.

Features

Code is the source of truth — Tokens live in TypeScript with full type-checking, review and history.
Figma variables generated — Native Figma variables produced through the API — not a plugin panel bolted on.
Drift is a CI failure — A stale consumer fails the build instead of surfacing as a visual bug in QA.
Multi-platform output — CSS, Tailwind, iOS, Android and JSON from one definition.
Conditional tokens — Values that vary by theme, platform or density — expressible because they live in code.
Incremental adoption — Generate alongside your existing library and diff before you cut over.

Design systems fail at the seam The tokens are not the hard part. The hard part is that Figma and code are two independent databases with no referential integrity between them, and they diverge the moment anyone is in a hurry. Northlight fixes the direction of authority. Code is the source of truth. Figma variables are generated from it. Drift is impossible by construction rather than by discipline. The pipeline One definition, five consumers, one CI check that fails the build if any consumer is stale. Why not the other direction Generating code from Figma sounds friendlier to designers and is…

Read more on UneedLists

Ledgerly

Amara Diallo — Tue, 04 Aug 2026 07:00:00 GMT

Revenue analytics that tells you what changed instead of showing forty charts

Ledgerly connects to your billing provider and answers one question every morning: what moved, and does it matter? Movements are ranked by how much they should change your plans, with the supporting detail one click away instead of spread across a dashboard nobody reads.

Features

Ranked movements — A short daily list ordered by estimated impact, not a wall of charts.
Cohort-aware baselines — Compares a cohort against its own history rather than a meaningless global average.
Noise floor — Suppresses changes that are statistically indistinguishable from randomness at your volume.
Read-only by design — Refuses billing keys that carry write scopes.
One-click drill-down — Every movement expands to the accounts and invoices behind it.
Seat-independent pricing — Priced on tracked revenue. Inviting your co-founder costs nothing.

Dashboards answer questions nobody asked The standard revenue dashboard shows MRR, churn, ARPU, LTV and a dozen more, all at once, all the time. It is comprehensive and almost useless, because it never tells you which of those numbers deserves your attention this week. Ledgerly is built around ranking rather than display. The daily digest Every morning you get a short list. Each item is a movement — an acceleration, a cohort behaving differently, a plan whose churn changed — ranked by estimated impact on the next two quarters. Most days the list is short. That is the feature. What is underneat…

Read more on UneedLists