Auggie CLI Signals

Counted for All practitioners.

Positive, 21

1. Mixed

   parasti, developer whose team moved off Augment

   Comment on: Show HN: Semble - Code search for agents that uses 98% fewer tokens than grepHacker News comment

   Says the team used Augment Code and its Context Engine at work, then switched to Claude Code, which parasti finds reads code less efficiently.

   Published May 18, 2026self-described: Team at work, moved from Augment Code to Claude Code

   Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

2. Positive

   RanHal, Hacker News commenter

   Comment on: Ask HN: What do you still do manually in 2026 that should be automated?Hacker News comment

   Opened an Obsidian notes directory with Augment Code, which indexed it in seconds and made the notes retrievable.

   Published May 7, 2026self-described: Personal knowledge base

   Practitioner account, pseudonymous; direct use; counts toward Adoption and upkeep

3. Positive

   Kyle Getty, software engineer at Tilt

   AI DevOps: Building Intelligent Incident Response SystemsCompany engineering blog post, Tilt

   Getty describes an incident-response system as a real-world example: an alert triggers an Azure DevOps pipeline whose agent, built on the Auggie CLI over Agent Client Protocol, pulls the runbook, queries logs, reviews commits and posts a summary to Slack, using the Augment Context Engine for code analysis.

   "…with Agent Client Protocol we can build a C# based model to interact with the Augment CLI Auggie This allows us to leverage the benefits of the Augment Context Engine when doing things like code analysis."

   Published Feb 13, 2026Tilt on-call engineering on Azure, alert-triggered incident analysis

   Practitioner account, name and role on their organization's own page; ran the rollout; counts toward Delivered outcomes

4. Positive

   kenforthewin, Hacker News commenter, self-described

   Comment on: Launch HN: Nia (YC S25) - Give better context to coding agentsHacker News comment

   Says Augment's main value add is its context engine, which "imo" they do really well, and that if all they had to do was launch an MCP for it, the comparison is still worth exploring.

   Published Dec 8, 2025

   Practitioner account, pseudonymous; opinion; not counted, repeats a stronger account

5. Positive

   Codewello, developer and YouTube creator

   Roo Code vs. Copilot vs. Cline vs. Kilo Code vs. Augment Code: The Ultimate Coding Assistant.Video, YouTube

   Rates Augment the most polished VS Code assistant with one of the best context engines, but calls the US$60 plan too much and reports paying over US$150 in a month.

   "…augment code that will cover everything from $10 affordable plan to $60 or even going higher like I did last month and pay more than $150 per month for a coding assistant. So, choose your B and let me know down below what you think…"

   Published Oct 27, 2025self-described: Compares five VS Code assistants the channel uses

   Practitioner account, self-described; direct use; counts toward Cost predictability and limits, Adoption and upkeep

6. Positive

   Robert Matsuoka, fractional CTO and former CTO of Tripadvisor

   Augment Code's Auggie: When Focused Single-Process Beats Multi-AgentNewsletter post, Hyperdev (Substack)

   After Claude Code failed across several sessions to fix a D3.js visualization in his framework, Robert Matsuoka got it working with Auggie in about 45 minutes, crediting the Context Engine's automatic indexing.

   "Claude Code couldn't crack it across multiple sessions. The complexity of coordinating D3.js rendering with file system analysis kept leading us in circles."

   "Within 45 minutes on that train ride, Auggie and I had it working."

   "The Context Engine's automatic codebase understanding is genuinely superior to anything else I've used."

   Published Sep 4, 2025self-described: Solo, debugging his open-source Claude MPM framework, a long-time Claude Code user

   Practitioner account, self-described; direct use; counts toward Delivered outcomes, Reliability, Operational control, Cost predictability and limits

7. Positive

   David Eastman, London-based software developer and consultant, writing for The New Stack

   Developer Walk-Through of Auggie CLI, an Agentic Terminal AppIndependent review article, The New Stack

   Testing the Auggie CLI beta on a Rails project, David Eastman found it worked harder than other agentic CLIs to understand the project. But it showed no usage or cost and did not say which model it used.

   "What I did like was the sense that Augment is working a little harder to figure out your project, and how it works."

   "There has been no attempt to tell me what costs have been used, either from Augment or via the model providers. In fact, I still don’t know what model was ever in use…"

   Published Aug 23, 2025self-described: Solo, trying the private beta on his own Ruby on Rails tool for a game

   Practitioner account, self-described; direct use; counts toward Delivered outcomes, Operational control, Cost predictability and limits, Adoption and upkeep

8. Mixed

   Ricky Guo, daily Augment user

   Augment Code releases CLI Claude Code competitor Auggie (reply)Product Hunt forum reply

   Uses Augment daily after quitting Cursor and Windsurf and rates its context engineering top-tier, but recent versions crashed until a rollback to the previous week's release.

   Published Jul 31, 2025self-described: Moved from Cursor and Windsurf in spring 2025; uses VS Code

   Practitioner account, pseudonymous; direct use; counts toward Delivered outcomes, Reliability

9. Positive

   Allan Leinwand, Chief Technology Officer at Webflow

   Building with AI: How Webflow engineers move faster, think deeper, and ship smarterCompany engineering blog post, Webflow Blog

   Leinwand describes how Webflow uses Augment Code in VS Code, WebStorm or Cursor to answer long-context questions and find logic flaws, keeps Augment guidelines versioned in the codebase, and pairs remote agents with Cursor and Augment so engineers can work on many PRs at once.

   "We also leverage Augment Code to bring the best of LLM assisted engineering into VSCode, WebStorm or Cursor. It answers long-context questions, finds bad assumptions developers are making, helps find logic flaws, scaffolds unit tests, highlights regressions, and more."

   "These tools are only as good as the context you provide them which is why we've invested deeply in maintaining Cursor rules and Augment guidelines versioned within the codebase."

   Published Jul 16, 2025Webflow engineering, where every engineer gets Augment Code beside Cursor

   Practitioner account, self-described; opinion; not counted, takes no position

10. Mixed

    M4v3R, developer using Augment at a day job and on a side project

    Comment on: Show HN: I AI-coded a tower defense game and documented the whole processHacker News comment

    Augment's context engine worked on a medium-sized day-job codebase where other agents failed, but it costs more than Cursor and often times out.

    Published Jul 4, 2025self-described: Medium-sized codebase at a day job

    Practitioner account, pseudonymous; direct use; counts toward Cost predictability and limits

11. Positive

    mythz, Hacker News commenter

    Comment on: Claude Code is now available to Pro plansHacker News comment

    Moved to Augment Code for its monthly cap and context engine, which lets agents match existing implementations without context in every message.

    Published Jun 4, 2025self-described: Uses VS Code and Rider

    Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

12. Mixed

    hn_acc1, developer whose company allows only Augment

    Comment on: Human coders are still better than LLMsHacker News comment

    At a company that allows only Augment, finds its completions follow team debugging idioms, but it keeps suggesting a temporary hack from months earlier.

    Published May 30, 2025self-described: Company standard: the only AI tool allowed

    Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

13. Positive

    nchmy, Hacker News commenter, self-described

    Comment on: Why Cline doesn't index your codebaseHacker News comment

    Calls Augment Code the best coding agent around and credits its code indexer.

    Published May 28, 2025

    Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

14. Positive

    jjani, Hacker News commenter, self-described

    Comment on: Why Cline doesn't index your codebaseHacker News comment

    Agreed with the case against codebase indexing until trying Augment, whose indexing and retrieval changed that opinion.

    Published May 27, 2025

    Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

15. Positive

    cheema33, Hacker News commenter

    Comment on: The unreasonable effectiveness of an LLM agent loop with tool useHacker News comment

    Has Augment Code write a PRD from its index of local code, then feeds it one task at a time from a plan, with another model reviewing each task in git.

    Published May 20, 2025self-described: Works in a git repo with several LLMs reviewing each other

    Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

16. Mixed

    Kent Beck, software developer and author

    My Augmented Coding Tools As Of 16 May 2025Newsletter, Software Design: Tidy First? (Substack)

    Augment builds its own context from large projects and has a professional UI, but its habit of deleting tests is infuriating and trust-destroying.

    Published May 16, 2025self-described: Personal projects in VS Code and IntelliJ

    Practitioner account, self-described; direct use; counts toward Delivered outcomes

17. Positive

    bv_dev, Hacker News commenter

    Comment on: My 600 Hours with AI Coding Assistants: A Practical ComparisonHacker News comment

    After 600 hours across assistants, names Augment Code the go-to tool for multi-file refactors and features despite its higher price, noting occasional hallucinations outside the codebase.

    Published May 14, 2025self-described: Over 600 hours across AI coding assistants in three months

    Practitioner account, pseudonymous; direct use; counts toward Delivered outcomes, Cost predictability and limits

18. Positive

    evelant, Hacker News commenter, self-described

    Comment on: Cursor hits US$9B valuationHacker News comment

    Says Augment has no model picker and runs on Claude 3.7, and that its context engine is far ahead of other tools and usually gets it right where others fail.

    Published May 5, 2025

    Practitioner account, pseudonymous; opinion; not counted, repeats a stronger account

19. Mixed

    JonatanRoo, GitHub user

    Sub-agents bypass CLI permission enforcement (security gap) - v0.14.0GitHub issue on augmentcode/auggie

    A developer using Auggie for all work found that permission rules in settings.json blocked commands such as rm for the main agent, but sub-agents ran the same commands anyway.

    "CLI permission rules configured in settings.json work correctly for the main agent but are completely ignored by sub-agents."

    Published Jan 30, 2026self-described: Solo, using Auggie for all development for several months

    Practitioner account, pseudonymous; structured test; counts toward Operational control

20. Positive

    chime, Hacker News commenter

    Comment on: Why Cline doesn't index your codebaseHacker News comment

    Says Augment's indexing makes a large difference in a large monorepo, enough that chime would not use an editor without full indexing.

    Published May 27, 2025self-described: Large monorepo

    Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

21. Positive

    Josh Pocock, founder of Executive Stride and YouTube creator

    FREE Augment Code Beats Cursor & Windsurf For Large Codebase Indexing!Video, YouTube

    Indexed his Next.js codebase quickly and built a case-study feature with Augment, recommending it for existing projects that outgrow Cursor or Windsurf, with less benefit for new ones.

    "…the blog Pages or case studies first let's create the necessary interfaces and functions okay so we'll just see that it's doing that here now let's create the case studies list page create a new file so it can create new files…"

    Published Feb 22, 2025self-described: His company's Next.js site

    Practitioner account, self-described; direct use; counts toward Delivered outcomes, Operational control, Cost predictability and limits, Adoption and upkeep

Mixed or negative, 1

1. Mixed

   Povilas Korop, Laravel developer and creator of AI Coding Daily

   Augment Code: Worth US$50/Month? Three Tests vs Claude CodeVideo, YouTube

   On a relationship refactor, Augment took 16 minutes because it wrote and ran tests unprompted while Claude Code took 4 without tests, and Korop found no context-speed edge.

   "…and then fixed again, all of that in total took 12 extra minutes. So for the whole task, cloud code took 4 minutes, but without tests, augment took 16 minutes, but we have automated tests executed and verified as a result. It's your preference whether…"

   Published Aug 18, 2025self-described: Laravel project, three tasks run side by side in Augment and Claude Code

   Practitioner account, self-described; structured test; counts toward Delivered outcomes, Reliability, Cost predictability and limits

Counted for All practitioners.

1. No opinion stated

   Pengrui Lu and five co-authors, academic researchers

   ProjDevBench: Benchmarking AI Coding Agents on End-to-End Project DevelopmentIndependent benchmark, research paper on arXiv

   On ProjDevBench, the authors report Codex with GPT-5 highest at 77.85, outperforming Augment with GPT-5 at 72.35, Cursor with GPT-5 at 71.85 and Claude Code with Sonnet-4.5 at 68.87. Augment with Sonnet-4.5 scored 70.10.

   "Codex paired with GPT-5 achieves the highest final weighted score of 77.85, outperforming Augment+GPT-5 at 72.35, Cursor+GPT-5 at 71.85, and Claude Code+Sonnet-4.5 at 68.87."

   "Both Cursor and Augment demonstrate stable performance across different base models, with all their tested configurations achieving final scores above 70…"

   Published Feb 2, 2026self-described: Independent benchmark of six coding agents through their command-line interfaces on 20 project-building tasks

   Practitioner account; structured test; counts toward Delivered outcomes

2. Positive

   Nathan Sebhastian, developer and YouTube creator

   Auggie CLI + Sonnet 4: The BEST Agentic Terminal I've Tried Yet (Augment Code CLI)Video, YouTube

   Builds and pushes a CRM dashboard with Auggie CLI and its native GitHub integration, calling it a strong addition for Augment users at one of the highest prices.

   "What's interesting is that this native integration also works in Auggie, which means we can instruct Auggie to commit, push, create pull requestes, basically anything GitHub allows the agent to do. Let's see a quick demo. Uh here I will try to push the CRM…"

   Published Aug 29, 2025self-described: Builds a CRM dashboard in Auggie CLI

   Practitioner account, self-described; direct use; counts toward Operational control, Cost predictability and limits, Adoption and upkeep

3. Mixed

   Povilas Korop, Laravel developer and creator of AI Coding Daily

   Augment Code: Worth US$50/Month? Three Tests vs Claude CodeVideo, YouTube

   On a relationship refactor, Augment took 16 minutes because it wrote and ran tests unprompted while Claude Code took 4 without tests, and Korop found no context-speed edge.

   "…and then fixed again, all of that in total took 12 extra minutes. So for the whole task, cloud code took 4 minutes, but without tests, augment took 16 minutes, but we have automated tests executed and verified as a result. It's your preference whether…"

   Published Aug 18, 2025self-described: Laravel project, three tasks run side by side in Augment and Claude Code

   Practitioner account, self-described; structured test; counts toward Delivered outcomes, Reliability, Cost predictability and limits

4. Positive

   lemiffe, Hacker News commenter

   Comment on: Vibe coding is the fast fashion industry of software engineeringHacker News comment

   Augment has produced production-level code in dbt, PHP and TypeScript codebases, creating and running tests, with several levels of review before merge.

   Published Aug 1, 2025self-described: dbt, PHP and TypeScript codebases

   Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

5. Positive

   gavinray, engineer at a company that rolled out Copilot and Augment

   Comment on: Ask HN: Where are the AI-driven profits or promotions?Hacker News comment

   Their company offers Copilot org-wide and Augment, and Augment has had the biggest impact there, with small features and many refactors done entirely through it.

   Published Jul 13, 2025self-described: Company provides Copilot and Augment

   not counted, repeats a stronger account

6. Positive

   bv_dev, Hacker News commenter

   Comment on: My 600 Hours with AI Coding Assistants: A Practical ComparisonHacker News comment

   After 600 hours across assistants, names Augment Code the go-to tool for multi-file refactors and features despite its higher price, noting occasional hallucinations outside the codebase.

   Published May 14, 2025self-described: Over 600 hours across AI coding assistants in three months

   Practitioner account, pseudonymous; direct use; counts toward Delivered outcomes, Cost predictability and limits

7. Mixed

   Matt Lewellyn, VP of Engineering at Level 12

   AI Agent-Driven Development with Augment CodeCompany engineering blog post, Level 12

   Lewellyn had Augment Agent build a weighted-calculation feature: one prompt produced a grid view, a form, views with permissions and Jinja templates, a 70% solution that didn't work right away. It skipped the project's common mixins, duplicated template markup, and added no tests, which the prompt hadn't asked for.

   "I found that Augment Agent had significant productivity benefits by generating functional code structures quickly, but was limited in pattern recognition, code organization, and testing…"

   "For all of the criticisms, bear in mind that the agent essentially developed a 70% solution, which is a huge productivity boost."

   Published Apr 11, 2025Level 12, a Louisville software consultancy building Python SaaS apps for clients

   Practitioner account, name and role on their organization's own page; direct use; counts toward Delivered outcomes

8. Positive

   Robert Matsuoka, fractional CTO and former CTO of Tripadvisor

   Augment Code's Auggie: When Focused Single-Process Beats Multi-AgentNewsletter post, Hyperdev (Substack)

   After Claude Code failed across several sessions to fix a D3.js visualization in his framework, Robert Matsuoka got it working with Auggie in about 45 minutes, crediting the Context Engine's automatic indexing.

   "Claude Code couldn't crack it across multiple sessions. The complexity of coordinating D3.js rendering with file system analysis kept leading us in circles."

   "Within 45 minutes on that train ride, Auggie and I had it working."

   "The Context Engine's automatic codebase understanding is genuinely superior to anything else I've used."

   Published Sep 4, 2025self-described: Solo, debugging his open-source Claude MPM framework, a long-time Claude Code user

   Practitioner account, self-described; direct use; counts toward Delivered outcomes, Reliability, Operational control, Cost predictability and limits

9. Mixed

   Jessica Kerr (Jessitron), software developer

   Augmented Coding: an Experience ReportPersonal blog post, jessitron.com

   Got further than alone with Augment's agent, but it loops, forgets what is not in its memories, and needs smaller steps and end-to-end checks after every change.

   Published Apr 20, 2025self-described: Weekend project: an app implemented in three to four languages

   Practitioner account; direct use; counts toward Delivered outcomes, Reliability

Counted for All practitioners.

Positive, 8

1. Positive

   atombender, Hacker News commenter

   Comment on: Claude Code: Now in Beta in ZedHacker News comment

   Rates Augment's completions and agent highly after more than a year in JetBrains IDEs, and says colleagues moved from Cursor to Augment.

   Published Sep 3, 2025self-described: Uses Augment in JetBrains IDEs; colleagues use Cursor

   Practitioner account, pseudonymous; direct use; not counted, takes no position

2. Positive

   chime, Hacker News commenter

   Comment on: Ask HN: What useful AI tools do you use every day?Hacker News comment

   After trying most AI editors and CLI agents, reports the best results with the Augment extension in VS Code, including its remote agent.

   Published Jun 25, 2025self-described: Has tried most AI editors and CLI agents

   Practitioner account, pseudonymous; direct use; not counted, takes no position

3. Positive

   Matt Lewellyn, VP of Engineering at Level 12

   Building UI with AICompany blog post, Level 12

   Level 12's VP of Engineering built a client's user deletion dialog with Augment Code in VS Code by splitting the work into small prompts. He estimates it cut development time by 50 to 60%, though one revision went the wrong way and testing still took real time.

   "In my latest experiment, I used Augment Code in VSCode to construct a modal dialog flow."

   "Overall, using these prompts likely cut the development time by 50-60%. Testing, revision, and refinement remain a time-consuming and integral part of the development process."

   Published Apr 30, 2025Level 12, a Louisville software consultancy, on a client project with a Bootstrap template

   Practitioner account, name and role on their organization's own page; direct use; counts toward Delivered outcomes

4. Positive

   ilrwbwrkhv, Hacker News commenter, self-described

   Comment on: GitHub Copilot: The Agent AwakensHacker News comment

   Recommends trying Augment, saying its IntelliJ integration works really well and wondering why it isn't more widely known.

   Published Feb 6, 2025

   Practitioner account, pseudonymous; direct use; not counted, takes no position

5. Mixed

   atombender, Hacker News commenter

   Comment on: Claude Code gets native LSP supportHacker News comment

   Enjoys Augment's fast completions but uses Claude Code as the agent, and reports serious stability problems, including the IDE using all cores while re-indexing.

   Published Dec 23, 2025self-described: Uses Augment completions with Claude Code for agent work

   Practitioner account, pseudonymous; direct use; counts toward Reliability

6. Positive

   Codewello, developer and YouTube creator

   Roo Code vs. Copilot vs. Cline vs. Kilo Code vs. Augment Code: The Ultimate Coding Assistant.Video, YouTube

   Rates Augment the most polished VS Code assistant with one of the best context engines, but calls the US$60 plan too much and reports paying over US$150 in a month.

   "…augment code that will cover everything from $10 affordable plan to $60 or even going higher like I did last month and pay more than $150 per month for a coding assistant. So, choose your B and let me know down below what you think…"

   Published Oct 27, 2025self-described: Compares five VS Code assistants the channel uses

   Practitioner account, self-described; direct use; counts toward Cost predictability and limits, Adoption and upkeep

7. Positive

   Robert Matsuoka, fractional CTO and former CTO of Tripadvisor

   Augment Code's Auggie: When Focused Single-Process Beats Multi-AgentNewsletter post, Hyperdev (Substack)

   After Claude Code failed across several sessions to fix a D3.js visualization in his framework, Robert Matsuoka got it working with Auggie in about 45 minutes, crediting the Context Engine's automatic indexing.

   "Claude Code couldn't crack it across multiple sessions. The complexity of coordinating D3.js rendering with file system analysis kept leading us in circles."

   "Within 45 minutes on that train ride, Auggie and I had it working."

   "The Context Engine's automatic codebase understanding is genuinely superior to anything else I've used."

   Published Sep 4, 2025self-described: Solo, debugging his open-source Claude MPM framework, a long-time Claude Code user

   Practitioner account, self-described; direct use; counts toward Delivered outcomes, Reliability, Operational control, Cost predictability and limits

8. Mixed

   Jessica Kerr (Jessitron), software developer

   Augmented Coding: an Experience ReportPersonal blog post, jessitron.com

   Got further than alone with Augment's agent, but it loops, forgets what is not in its memories, and needs smaller steps and end-to-end checks after every change.

   Published Apr 20, 2025self-described: Weekend project: an app implemented in three to four languages

   Practitioner account; direct use; counts toward Delivered outcomes, Reliability

Mixed or negative, 5

1. Mixed

   vb-8448, Hacker News commenter

   Comment on: Claude Code gets native LSP supportHacker News comment

   Currently uses Augment and finds it good, but its completions trail Supermaven and Cursor and lack automatic cursor movement.

   Published Dec 22, 2025self-described: Uses PyCharm

   Practitioner account, pseudonymous; direct use; not counted, takes no position

2. Positive

   David Eastman, London-based software developer and consultant, writing for The New Stack

   Developer Walk-Through of Auggie CLI, an Agentic Terminal AppIndependent review article, The New Stack

   Testing the Auggie CLI beta on a Rails project, David Eastman found it worked harder than other agentic CLIs to understand the project. But it showed no usage or cost and did not say which model it used.

   "What I did like was the sense that Augment is working a little harder to figure out your project, and how it works."

   "There has been no attempt to tell me what costs have been used, either from Augment or via the model providers. In fact, I still don’t know what model was ever in use…"

   Published Aug 23, 2025self-described: Solo, trying the private beta on his own Ruby on Rails tool for a game

   Practitioner account, self-described; direct use; counts toward Delivered outcomes, Operational control, Cost predictability and limits, Adoption and upkeep

3. Mixed

   kwalitynut, member of a Clari developer-experience team

   [Feature]: Project-scoped hooks (.augment/settings.json checked into the repo)GitHub issue

   A team supporting both Augment and Claude Code asks for hooks a repository can ship, since Augment loads hooks only from user and system settings.

   Published May 1, 2026self-described: Team building an AI-provenance commit standard for Augment and Claude Code users

   Practitioner account, pseudonymous; direct use; counts toward Operational control

4. Mixed

   chuckadams, Hacker News commenter

   Comment on: Agentic Coding RecommendationsHacker News comment

   Liked Augment apart from its heavy UI, then went back to JetBrains Junie once Augment revealed its price.

   Published Jun 12, 2025self-described: Works in JetBrains IDEs

   Practitioner account, pseudonymous; direct use; not counted, takes no position

5. Mixed

   hn_acc1, developer whose company allows only Augment

   Comment on: Human coders are still better than LLMsHacker News comment

   At a company that allows only Augment, finds its completions follow team debugging idioms, but it keeps suggesting a temporary hack from months earlier.

   Published May 30, 2025self-described: Company standard: the only AI tool allowed

   Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

Counted for All practitioners.

Positive, 2

1. Positive

   chime, Hacker News commenter

   Comment on: Ask HN: What useful AI tools do you use every day?Hacker News comment

   After trying most AI editors and CLI agents, reports the best results with the Augment extension in VS Code, including its remote agent.

   Published Jun 25, 2025self-described: Has tried most AI editors and CLI agents

   Practitioner account, pseudonymous; direct use; not counted, takes no position

2. Mixed

   Kent Beck, software developer and author

   My Augmented Coding Tools As Of 16 May 2025Newsletter, Software Design: Tidy First? (Substack)

   Augment builds its own context from large projects and has a professional UI, but its habit of deleting tests is infuriating and trust-destroying.

   Published May 16, 2025self-described: Personal projects in VS Code and IntelliJ

   Practitioner account, self-described; direct use; counts toward Delivered outcomes

Mixed or negative, 4

1. Mixed

   fodurrr, GitHub user

   Predatory credit pricing: --print re-indexes workspace on every call, burning 2000+ credits per queryGitHub issue on augmentcode/auggie

   Scripting Auggie CLI in print mode, the author found every call re-indexed the workspace, and 17 queries in one day used 37,016 credits, 92.5% of the Indie plan's monthly allowance.

   "Burned **37,016 credits** (92.5% of 40,000 monthly Indie Plan allowance)…"

   Published Feb 20, 2026self-described: Solo, scripting Auggie CLI in print mode on the Indie plan

   Practitioner account, pseudonymous; direct use; counts toward Cost predictability and limits

2. Mixed

   M4v3R, developer using Augment at a day job and on a side project

   Comment on: Show HN: I AI-coded a tower defense game and documented the whole processHacker News comment

   Augment's context engine worked on a medium-sized day-job codebase where other agents failed, but it costs more than Cursor and often times out.

   Published Jul 4, 2025self-described: Medium-sized codebase at a day job

   Practitioner account, pseudonymous; direct use; counts toward Cost predictability and limits

3. Mixed

   chuckadams, Hacker News commenter

   Comment on: Agentic Coding RecommendationsHacker News comment

   Liked Augment apart from its heavy UI, then went back to JetBrains Junie once Augment revealed its price.

   Published Jun 12, 2025self-described: Works in JetBrains IDEs

   Practitioner account, pseudonymous; direct use; not counted, takes no position

4. Positive

   bv_dev, Hacker News commenter

   Comment on: My 600 Hours with AI Coding Assistants: A Practical ComparisonHacker News comment

   After 600 hours across assistants, names Augment Code the go-to tool for multi-file refactors and features despite its higher price, noting occasional hallucinations outside the codebase.

   Published May 14, 2025self-described: Over 600 hours across AI coding assistants in three months

   Practitioner account, pseudonymous; direct use; counts toward Delivered outcomes, Cost predictability and limits

Counted for All practitioners.

1. Mixed

   inercia, Auggie CLI user running it under an ACP host

   Multi-minute event-loop freeze in uv__fsevents_close (macOS) when indexing/file-watching is enabledGitHub issue

   Reports that with indexing on, Auggie's agent froze for about five to seven minutes on macOS before finishing, traces the hang to file watching, and saw it on several workspaces at once.

   Published Jun 4, 2026self-described: Large workspaces on macOS, Auggie 0.28.0 in ACP mode with indexing on

   Practitioner account, pseudonymous; direct use; counts toward Reliability

2. Mixed

   yann0001, Auggie CLI user, self-described

   [Bug]: 0.15.0 Context exceedsGitHub issue

   Reports that Auggie 0.15.0 often fails with a prompt-length error, a problem present since its first releases.

   Published Jan 30, 2026

   Practitioner account, pseudonymous; direct use; counts toward Reliability

3. Mixed

   atombender, Hacker News commenter

   Comment on: Claude Code gets native LSP supportHacker News comment

   Enjoys Augment's fast completions but uses Claude Code as the agent, and reports serious stability problems, including the IDE using all cores while re-indexing.

   Published Dec 23, 2025self-described: Uses Augment completions with Claude Code for agent work

   Practitioner account, pseudonymous; direct use; counts toward Reliability

4. Mixed

   rishitank, Auggie CLI user

   [Bug]: New version of Auggie (v0.6.0) having delayed responsesGitHub issue

   Reports that Auggie 0.6.0 sometimes takes a long time to answer a prompt with no sign that it is working.

   Published Oct 31, 2025self-described: macOS, zsh

   Practitioner account, pseudonymous; direct use; counts toward Reliability

5. Mixed

   Povilas Korop, Laravel developer and creator of AI Coding Daily

   Augment Code: Worth US$50/Month? Three Tests vs Claude CodeVideo, YouTube

   On a relationship refactor, Augment took 16 minutes because it wrote and ran tests unprompted while Claude Code took 4 without tests, and Korop found no context-speed edge.

   "…and then fixed again, all of that in total took 12 extra minutes. So for the whole task, cloud code took 4 minutes, but without tests, augment took 16 minutes, but we have automated tests executed and verified as a result. It's your preference whether…"

   Published Aug 18, 2025self-described: Laravel project, three tasks run side by side in Augment and Claude Code

   Practitioner account, self-described; structured test; counts toward Delivered outcomes, Reliability, Cost predictability and limits

6. Mixed

   Ricky Guo, daily Augment user

   Augment Code releases CLI Claude Code competitor Auggie (reply)Product Hunt forum reply

   Uses Augment daily after quitting Cursor and Windsurf and rates its context engineering top-tier, but recent versions crashed until a rollback to the previous week's release.

   Published Jul 31, 2025self-described: Moved from Cursor and Windsurf in spring 2025; uses VS Code

   Practitioner account, pseudonymous; direct use; counts toward Delivered outcomes, Reliability

7. Mixed

   physix, Hacker News commenter

   Comment on: Ask HN: What are your favorite coding tools?Hacker News comment

   Names the Augment Code plugin for IDEA and PyCharm a current favorite tool, joking that its waits leave time to read Hacker News.

   Published Jul 15, 2025self-described: Uses the plugin in IDEA and PyCharm

   Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

8. Mixed

   M4v3R, developer using Augment at a day job and on a side project

   Comment on: Show HN: I AI-coded a tower defense game and documented the whole processHacker News comment

   Augment's context engine worked on a medium-sized day-job codebase where other agents failed, but it costs more than Cursor and often times out.

   Published Jul 4, 2025self-described: Medium-sized codebase at a day job

   Practitioner account, pseudonymous; direct use; counts toward Cost predictability and limits

9. Positive

   parasti, developer on a team of four rewriting a legacy CRM

   Comment on: Ask HN: How is AI-assisted coding going for you professionally?Hacker News comment

   On a team of four using Augment Code heavily to rewrite a legacy CRM with 30 years of business rules, the author works in the Auggie CLI because the VS Code integration became too slow and hung often. A favorite recent interaction paired two sessions through a shared API specification.

   Published Mar 15, 2026self-described: Team of 4, full rewrite of a 30-year-old CRM in Laravel

   Practitioner account, pseudonymous; direct use; counts toward Operational control

Counted for All practitioners.

Raised as a concern, 7

1. Mixed

   fodurrr, GitHub user

   Predatory credit pricing: --print re-indexes workspace on every call, burning 2000+ credits per queryGitHub issue on augmentcode/auggie

   Scripting Auggie CLI in print mode, the author found every call re-indexed the workspace, and 17 queries in one day used 37,016 credits, 92.5% of the Indie plan's monthly allowance.

   "Burned **37,016 credits** (92.5% of 40,000 monthly Indie Plan allowance)…"

   Published Feb 20, 2026self-described: Solo, scripting Auggie CLI in print mode on the Indie plan

   Practitioner account, pseudonymous; direct use; counts toward Cost predictability and limits

2. Mixed

   mythz, Hacker News commenter

   Comment on: How Google got its groove back and edged ahead of OpenAIHacker News comment

   Retried Antigravity a few weeks after launch, after Augment Code's new pricing kicked in, and was pleasantly surprised how reliable it had become on the free quota. Upgraded to Pro and hasn't hit a quota since.

   Published Jan 8, 2026self-described: Full-time coding assistant user

   Practitioner account, pseudonymous; opinion; not counted, takes no position

3. Mixed

   feintruled, Hacker News commenter, self-described

   Comment on: Developers are choosing older AI modelsHacker News comment

   Notes Augment's switch to token-based pricing has not gone down well, stopped using GPT-5 in it, and finds Sonnet 4.5 too verbose for token billing.

   Published Nov 5, 2025

   Practitioner account, pseudonymous; direct use; counts toward Cost predictability and limits

4. Negative

   nchmy, Hacker News commenter, self-described

   Comment on: </> Htmx - The Fetch()eningHacker News comment

   Says the author has been rugpulled a few times recently by thoroughly unscrupulous companies, calling Augment Code "trash humans".

   Published Nov 3, 2025

   Practitioner account, pseudonymous; opinion; not counted, takes no position

5. Positive

   Codewello, developer and YouTube creator

   Roo Code vs. Copilot vs. Cline vs. Kilo Code vs. Augment Code: The Ultimate Coding Assistant.Video, YouTube

   Rates Augment the most polished VS Code assistant with one of the best context engines, but calls the US$60 plan too much and reports paying over US$150 in a month.

   "…augment code that will cover everything from $10 affordable plan to $60 or even going higher like I did last month and pay more than $150 per month for a coding assistant. So, choose your B and let me know down below what you think…"

   Published Oct 27, 2025self-described: Compares five VS Code assistants the channel uses

   Practitioner account, self-described; direct use; counts toward Cost predictability and limits, Adoption and upkeep

6. Mixed

   chuckadams, Hacker News commenter

   Comment on: Agentic Coding RecommendationsHacker News comment

   Liked Augment apart from its heavy UI, then went back to JetBrains Junie once Augment revealed its price.

   Published Jun 12, 2025self-described: Works in JetBrains IDEs

   Practitioner account, pseudonymous; direct use; not counted, takes no position

7. Positive

   David Eastman, London-based software developer and consultant, writing for The New Stack

   Developer Walk-Through of Auggie CLI, an Agentic Terminal AppIndependent review article, The New Stack

   Testing the Auggie CLI beta on a Rails project, David Eastman found it worked harder than other agentic CLIs to understand the project. But it showed no usage or cost and did not say which model it used.

   "What I did like was the sense that Augment is working a little harder to figure out your project, and how it works."

   "There has been no attempt to tell me what costs have been used, either from Augment or via the model providers. In fact, I still don’t know what model was ever in use…"

   Published Aug 23, 2025self-described: Solo, trying the private beta on his own Ruby on Rails tool for a game

   Practitioner account, self-described; direct use; counts toward Delivered outcomes, Operational control, Cost predictability and limits, Adoption and upkeep

Discussed without concern, 5

1. Positive

   Nathan Sebhastian, developer and YouTube creator

   Auggie CLI + Sonnet 4: The BEST Agentic Terminal I've Tried Yet (Augment Code CLI)Video, YouTube

   Builds and pushes a CRM dashboard with Auggie CLI and its native GitHub integration, calling it a strong addition for Augment users at one of the highest prices.

   "What's interesting is that this native integration also works in Auggie, which means we can instruct Auggie to commit, push, create pull requestes, basically anything GitHub allows the agent to do. Let's see a quick demo. Uh here I will try to push the CRM…"

   Published Aug 29, 2025self-described: Builds a CRM dashboard in Auggie CLI

   Practitioner account, self-described; direct use; counts toward Operational control, Cost predictability and limits, Adoption and upkeep

2. Mixed

   Povilas Korop, Laravel developer and creator of AI Coding Daily

   Augment Code: Worth US$50/Month? Three Tests vs Claude CodeVideo, YouTube

   On a relationship refactor, Augment took 16 minutes because it wrote and ran tests unprompted while Claude Code took 4 without tests, and Korop found no context-speed edge.

   "…and then fixed again, all of that in total took 12 extra minutes. So for the whole task, cloud code took 4 minutes, but without tests, augment took 16 minutes, but we have automated tests executed and verified as a result. It's your preference whether…"

   Published Aug 18, 2025self-described: Laravel project, three tasks run side by side in Augment and Claude Code

   Practitioner account, self-described; structured test; counts toward Delivered outcomes, Reliability, Cost predictability and limits

3. Mixed

   M4v3R, developer using Augment at a day job and on a side project

   Comment on: Show HN: I AI-coded a tower defense game and documented the whole processHacker News comment

   Augment's context engine worked on a medium-sized day-job codebase where other agents failed, but it costs more than Cursor and often times out.

   Published Jul 4, 2025self-described: Medium-sized codebase at a day job

   Practitioner account, pseudonymous; direct use; counts toward Cost predictability and limits

4. Positive

   mythz, Hacker News commenter

   Comment on: Claude Code is now available to Pro plansHacker News comment

   Moved to Augment Code for its monthly cap and context engine, which lets agents match existing implementations without context in every message.

   Published Jun 4, 2025self-described: Uses VS Code and Rider

   Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

5. Positive

   bv_dev, Hacker News commenter

   Comment on: My 600 Hours with AI Coding Assistants: A Practical ComparisonHacker News comment

   After 600 hours across assistants, names Augment Code the go-to tool for multi-file refactors and features despite its higher price, noting occasional hallucinations outside the codebase.

   Published May 14, 2025self-described: Over 600 hours across AI coding assistants in three months

   Practitioner account, pseudonymous; direct use; counts toward Delivered outcomes, Cost predictability and limits

GILD's fact checks, which don't count as sources

Checked Oct 12, 2026.Standard costs US$20 and Business US$100 per month, each a flat price for up to 50 seats. Each includes the same dollar amount of pooled usage, billed at model list price plus a 40% service fee. Per-user budgets are Enterprise-only. A report about the 2025 message or credit plans does not describe these terms. Token-Based Pricing

Counted for All practitioners.

1. Negative

   hn_acc1, developer whose company standardizes on Augment

   Comment on: OpenAI raises US$110B on US$730B pre-money valuationHacker News comment

   Cannot get Augment with Opus 4.5 to edit a few C++ files in VS Code without it looping between the same investigation steps.

   Published Feb 27, 2026self-described: C++ work in VS Code

   Practitioner account, pseudonymous; direct use; counts toward Reliability

2. Negative

   danielodievich, engineer working with the company's support team

   Comment on: OpenAI declares 'code red' as Google catches up in AI raceHacker News comment

   When a colleague ran Augment with the feature's full source code to answer a customer question, it confidently made up behavior that does not exist.

   Published Dec 2, 2025self-described: Company support and engineering, answering a customer question

   Practitioner account, pseudonymous; second-hand; not counted, repeats a stronger account

3. Mixed

   replwoacause, Hacker News commenter

   Comment on: Cline is quietly eating Cursor's lunch and changing how we codeHacker News comment

   Pays for Augment Code alongside Copilot, finds it decent for simple work, and prefers Cline for large undertakings.

   Published May 8, 2025self-described: Pays for several assistants at once

   Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

4. Mixed

   Jessica Kerr (Jessitron), software developer

   Augmented Coding: an Experience ReportPersonal blog post, jessitron.com

   Got further than alone with Augment's agent, but it loops, forgets what is not in its memories, and needs smaller steps and end-to-end checks after every change.

   Published Apr 20, 2025self-described: Weekend project: an app implemented in three to four languages

   Practitioner account; direct use; counts toward Delivered outcomes, Reliability

5. Mixed

   Kent Beck, software developer and author

   My Augmented Coding Tools As Of 16 May 2025Newsletter, Software Design: Tidy First? (Substack)

   Augment builds its own context from large projects and has a professional UI, but its habit of deleting tests is infuriating and trust-destroying.

   Published May 16, 2025self-described: Personal projects in VS Code and IntelliJ

   Practitioner account, self-described; direct use; counts toward Delivered outcomes

6. Mixed

   Matt Lewellyn, VP of Engineering at Level 12

   AI Agent-Driven Development with Augment CodeCompany engineering blog post, Level 12

   Lewellyn had Augment Agent build a weighted-calculation feature: one prompt produced a grid view, a form, views with permissions and Jinja templates, a 70% solution that didn't work right away. It skipped the project's common mixins, duplicated template markup, and added no tests, which the prompt hadn't asked for.

   "I found that Augment Agent had significant productivity benefits by generating functional code structures quickly, but was limited in pattern recognition, code organization, and testing…"

   "For all of the criticisms, bear in mind that the agent essentially developed a 70% solution, which is a huge productivity boost."

   Published Apr 11, 2025Level 12, a Louisville software consultancy building Python SaaS apps for clients

   Practitioner account, name and role on their organization's own page; direct use; counts toward Delivered outcomes

Counted for All practitioners.

Raised as a concern, 5

1. Negative

   danielodievich, engineer working with the company's support team

   Comment on: OpenAI declares 'code red' as Google catches up in AI raceHacker News comment

   When a colleague ran Augment with the feature's full source code to answer a customer question, it confidently made up behavior that does not exist.

   Published Dec 2, 2025self-described: Company support and engineering, answering a customer question

   Practitioner account, pseudonymous; second-hand; not counted, repeats a stronger account

2. Positive

   Paul Schleger, Cyoda

   Getting the Most out of Augment CodeEngineering team write-up, Substack (Paul Schleger)

   Cyoda's team treats Augment as a skill: guidelines and memories, small checkpointed tasks, supervised diffs, and critical review of generated tests and docs.

   Published Sep 20, 2025self-described: Cyoda engineering team, six months of daily use, mostly IntelliJ IDEA with Java and Kotlin

   Practitioner account, self-described; ran the rollout; counts toward Delivered outcomes, Operational control, Adoption and upkeep

3. Mixed

   Kent Beck, software developer and author

   My Augmented Coding Tools As Of 16 May 2025Newsletter, Software Design: Tidy First? (Substack)

   Augment builds its own context from large projects and has a professional UI, but its habit of deleting tests is infuriating and trust-destroying.

   Published May 16, 2025self-described: Personal projects in VS Code and IntelliJ

   Practitioner account, self-described; direct use; counts toward Delivered outcomes

4. Mixed

   Jessica Kerr (Jessitron), software developer

   Augmented Coding: an Experience ReportPersonal blog post, jessitron.com

   Got further than alone with Augment's agent, but it loops, forgets what is not in its memories, and needs smaller steps and end-to-end checks after every change.

   Published Apr 20, 2025self-described: Weekend project: an app implemented in three to four languages

   Practitioner account; direct use; counts toward Delivered outcomes, Reliability

5. Mixed

   Matt Lewellyn, VP of Engineering at Level 12

   AI Agent-Driven Development with Augment CodeCompany engineering blog post, Level 12

   Lewellyn had Augment Agent build a weighted-calculation feature: one prompt produced a grid view, a form, views with permissions and Jinja templates, a 70% solution that didn't work right away. It skipped the project's common mixins, duplicated template markup, and added no tests, which the prompt hadn't asked for.

   "I found that Augment Agent had significant productivity benefits by generating functional code structures quickly, but was limited in pattern recognition, code organization, and testing…"

   "For all of the criticisms, bear in mind that the agent essentially developed a 70% solution, which is a huge productivity boost."

   Published Apr 11, 2025Level 12, a Louisville software consultancy building Python SaaS apps for clients

   Practitioner account, name and role on their organization's own page; direct use; counts toward Delivered outcomes

Discussed without concern, 6

1. Positive

   rsyring, Hacker News commenter

   Comment on: Task Paralysis and AIHacker News comment

   Has had GPT 5.4 write code and Opus review it in a fresh Augment session for eight months, with a spec file per change and one AGENTS.md for instructions.

   Published May 10, 2026self-described: Augment agents in VS Code and the CLI for about eight months

   Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

2. Positive

   Allan Leinwand, Chief Technology Officer at Webflow

   How Webflow Got 89% of Its Engineers To Use AI DailyInterview article, The New Stack

   In an interview, Webflow's CTO reports that Webflow has enabled agents in both Cursor and Augment Code, and engineers hand Augment Code a specific issue and get back suggested pull request changes that a human then reviews and tests. He describes engineers using MCP from Augment Code to reach Jira, GitHub and parts of the codebase.

   "…but we definitely have the ability to take a piece of code, put it up into Augment Code, and say: Go work on this particular issue…"

   "If I’m in Cursor and I’m in Augment Code, I use MCP to talk to Jira, I use MCP to talk to GitHub…"

   Published Aug 29, 2025Webflow engineering, about 300 engineers with Cursor and Augment Code in a company AI toolkit

   not counted until substantiated

3. Positive

   lemiffe, Hacker News commenter

   Comment on: Vibe coding is the fast fashion industry of software engineeringHacker News comment

   Augment has produced production-level code in dbt, PHP and TypeScript codebases, creating and running tests, with several levels of review before merge.

   Published Aug 1, 2025self-described: dbt, PHP and TypeScript codebases

   Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

4. Positive

   gondo, Hacker News commenter

   Comment on: Cursor 1.0Hacker News comment

   Reports that AugmentCode has done most of the coding for two weeks, with gondo prompting, reviewing and re-prompting across two or three parallel tasks in JetBrains IDEs.

   Published Jun 5, 2025self-described: Works in GoLand, Android Studio and other JetBrains IDEs

   Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

5. Positive

   cheema33, Hacker News commenter

   Comment on: The unreasonable effectiveness of an LLM agent loop with tool useHacker News comment

   Has Augment Code write a PRD from its index of local code, then feeds it one task at a time from a plan, with another model reviewing each task in git.

   Published May 20, 2025self-described: Works in a git repo with several LLMs reviewing each other

   Practitioner account, pseudonymous; direct use; not counted, repeats a stronger account

6. Positive

   Matt Lewellyn, VP of Engineering at Level 12

   Building UI with AICompany blog post, Level 12

   Level 12's VP of Engineering built a client's user deletion dialog with Augment Code in VS Code by splitting the work into small prompts. He estimates it cut development time by 50 to 60%, though one revision went the wrong way and testing still took real time.

   "In my latest experiment, I used Augment Code in VSCode to construct a modal dialog flow."

   "Overall, using these prompts likely cut the development time by 50-60%. Testing, revision, and refinement remain a time-consuming and integral part of the development process."

   Published Apr 30, 2025Level 12, a Louisville software consultancy, on a client project with a Bootstrap template

   Practitioner account, name and role on their organization's own page; direct use; counts toward Delivered outcomes