Palmier

What do they actually do

Palmier runs hosted “coding agents” against your GitHub repositories to do hands-on engineering tasks. After signing in with GitHub, you select a repo/branch and trigger agents from a web dashboard or by mentioning a GitHub bot on issues/PRs. The agents can open pull requests, add commits, summarize or review PRs, and answer questions about the codebase; activity and results are tracked on a Runs page and you can wire up Slack notifications and triggers Palmier homepage, Docs.

The product is live and self-serve, with a free Hobby tier and paid Pro, Teams, and Enterprise plans; limits and pricing are public. Enterprise options include self‑hosting and SLAs Pricing, Palmier homepage.

Practically, agents operate within finite context windows. Palmier uses Projects and shared context documents to preserve relevant context across multi‑step tasks; some deeper CI/CD troubleshooting features are noted as “coming soon,” not generally available today Docs, YC profile/launch.

Who are their target customer(s)

• Small startup engineering teams (2–10 devs): Too much time goes to small PRs, manual reviews, and triage, which slows feature work and forces context switching.
• Individual open‑source maintainers or solo developers: High issue and PR volume overwhelms limited time; reviews and fixes pile up and stall contributions.
• Product engineering teams automating repetitive workflows: Repetitive tasks like PR descriptions, changelogs, and standard fixes consume hours that could go to higher‑value work.
• DevOps/CI teams handling failing pipelines: Diagnosing CI failures across logs and repos takes deep context and delays releases; recurring build issues waste on‑call time.
• Engineering managers with review backlogs: Slow or inconsistent reviews create merge delays and quality regressions; they want standardized reviews and faster time‑to‑merge.

How would they acquire their first 10, 50, and 100 customers

• First 10: Run high‑touch pilots from YC/founder networks, early Product Hunt responders, and OSS maintainers; offer free/sponsored accounts and 1:1 onboarding to connect repos and ship a first agent‑generated PR Docs, Pricing, YC profile/launch, Product Hunt.
• First 50: Do targeted community outreach (OSS, Slack groups, HN), publish automation templates and public example PRs via the GitHub bot, and convert short paid trials into Pro/Teams with referral credits Palmier homepage, Docs, Pricing.
• First 100: Scale self‑serve (GitHub auth, clear limits, tutorials), publish case studies, and partner with adjacent tools/agencies to co‑sell CI/CD and issue‑tracker workflows; add referral credits and seed enterprise pilots Docs, Pricing, YC profile/launch.

What is the rough total addressable market

Top-down context:

There are ~27M professional developers globally (2024), and industry surveys indicate many work in small teams; GitHub’s platform reach (100M+ developers) underscores broad distribution potential Evans Data, JetBrains, Developer Nation, GitHub blog.

Bottom-up calculation:

Starting from 27M developers, assume 50% are in small teams and an average team size of 5 → ~2.7M teams; if each pays ~$2,940/year (5 × $49 × 12), the theoretical ARR is ~$7.94B Evans Data, JetBrains, Developer Nation, Pricing.

Assumptions:

• 50% of professional developers work in small teams likely to adopt repo‑level agents.
• Average small‑team size is 5 developers.
• Teams purchase the Teams plan at list price ($49/user/month).

Who are some of their notable competitors

• GitHub Copilot: In‑IDE AI with chat and an agent mode that can be assigned issues, make code changes, and create PRs; overlaps on repo‑level automation and GitHub‑native workflows, while often used as a real‑time pair programmer rather than a separate runs dashboard Copilot features, Coding agent.
• Sourcegraph Cody: AI assistant layered on Sourcegraph’s code search and indexing for large repos; overlaps on multi‑file edits and repo‑wide help, with a search‑first approach vs. a hosted agent dashboard and Slack/GitHub automations Cody guide, Sourcegraph overview.
• Sourcery: Automated code review and refactoring (notably strong in Python) that runs on PRs and can open suggested fix PRs; focuses on refactors and code‑quality checks rather than multi‑agent orchestration Sourcery product, GitHub integration.
• Replit Ghostwriter / Replit Agent: AI inside a cloud IDE that can generate/edit multi‑file apps, iterate, and deploy; overlaps on coordinated edits, but Replit centers on an in‑browser IDE experience vs. running agents against existing repos and team workflows Replit Agent docs.
• Codeium: IDE‑centric assistant for fast completions, chat, and repo search with enterprise deployment options; overlaps on speeding dev workflows but is primarily an in‑editor assistant vs. an autonomous bot/dashboard that opens PRs Codeium.