Archive
Tags
193 tags across 93 posts.
Most used
- OneCamp 88
- OpenSource 80
- Self-Hosted 80
- Go 38
- AI 23
- NextJS 21
- Architecture 15
- Project Management 13
- Security 11
- Governance 10
- Agents 9
- AI Agents 7
- Agencies 7
- Slack Alternative 7
- UX 6
- Kanban 5
- Reliability 5
- Audit 4
- Dgraph 4
- Docker 4
- MCP 4
- PostgreSQL 4
- Postgres 4
- Product 4
- Productivity 4
- Project 4
- Release 4
- Testing 4
- Time Tracking 4
- Chat 3
- Client Portal 3
- Gantt 3
- GitHub 3
- Notion Alternative 3
- Ollama 3
- OpenSearch 3
Everything, A to Z
- 2FA 1
- A2A 2
- AG-UI 1
- AGPL 1
- AI 23
- AI Agents 7
- AI Governance 2
- API 2
- API-Resilience 1
- Accessibility 1
- Agencies 7
- Agents 9
- Airtable Alternative 2
- Apps 1
- Architecture 15
- Asana 1
- Async Standups 1
- Audit 4
- Automation 1
- Backups 1
- Blockchain 2
- Burndown 1
- CI 1
- Changelog 1
- Charts 1
- Chat 3
- Check-ins 1
- CircuitBreaker 1
- ClamAV 1
- ClickUp 1
- Client Portal 3
- Clips 1
- CodePR 2
- CodeQuality 1
- Comparison 1
- Connectors 1
- Cost 1
- Cumulative Flow 1
- Custom Fields 2
- Dashboards 1
- Data-Migration 1
- DataSources 1
- Database 1
- DeadCode 1
- Debugging 2
- Decompression 1
- Defaults 1
- Delegation 1
- Dependencies 2
- Deployment 1
- Design 2
- Desktop App 1
- DevOps 2
- Dgraph 4
- Distributed Systems 1
- Docker 4
- DockerCompose 1
- Docs 1
- Documents 1
- Draw 1
- Durable 1
- EU AI Act 2
- Elicitation 1
- Email 1
- Enterprise 2
- Ethereum 2
- Excalidraw 1
- Focus Time 1
- Formulas 1
- Free Plan 1
- Freelancers 1
- Gantt 3
- GitHub 3
- Gmail 2
- Go 38
- Goals 1
- Governance 10
- GraphDB 1
- GraphRAG 1
- Guests 1
- Import 2
- Imports 1
- IndieHacker 1
- Interop 1
- Invoices 1
- Invoicing 2
- Jira 1
- Jira Alternative 1
- Kanban 5
- Keyboard Shortcuts 1
- LLM 2
- LiveKit 1
- MCP 4
- MQTT 2
- Managed Hosting 1
- Marketplace 1
- Mattermost 1
- Meetings 1
- Mentions 1
- Migration 1
- Migrations 1
- Mobile 1
- Moderation 1
- Multiplayer 1
- NextJS 21
- NodeJS 1
- Notion Alternative 3
- Nudges 1
- OAuth 2
- OKR 1
- OVH 1
- Observability 1
- Ollama 3
- Onboarding 2
- OneCamp 88
- Open Source 2
- OpenSearch 3
- OpenSource 80
- Operations 2
- Passkeys 1
- Performance 2
- Permissions 1
- Post-Mortem 2
- PostgreSQL 4
- Postgres 4
- Product 4
- Productivity 4
- Profiles 1
- Project 4
- Project Management 13
- Provisioning 1
- Python 1
- Read Receipts 1
- Real-Time 1
- Realtime 2
- Redis 1
- Redux 1
- Relations 1
- Release 4
- Reliability 5
- Remote Work 1
- Reports 2
- Rollups 1
- SCIM 2
- SDK 1
- SQL 1
- SSE 1
- SaaS 2
- Sandbox 2
- Scheduling 1
- Security 11
- Self-Hosted 80
- SelfHosted 2
- Shell 1
- Silent Failure 2
- Single Sign-On 1
- Slack 2
- Slack Alternative 7
- Slash-Commands 1
- Split View 1
- Sprints 1
- Startup 1
- StaticAnalysis 1
- TLS 1
- TOTP 1
- Tables 2
- Tasks 1
- Teams Alternative 1
- Templates 2
- Testing 4
- Time Tracking 4
- Timeline 2
- Tool-Calling 1
- Traefik 1
- Transcription 1
- TypeScript 2
- UTF-8 1
- UX 6
- Vision 1
- Voice 1
- Voice Notes 1
- Web-Performance 1
- WebRTC 1
- WebSockets 1
- Webhooks 1
- Websocket 1
- Whiteboard 1
- Workflows 1
- Workload 2
- XSS 1
- Yjs 1
- Zero-Trust 1
- monday.com 1
2FA 1 post
A2A 2 posts
AG-UI 1 post
AGPL 1 post
AI 23 posts
- Describe a project in a sentence, start with its plan
- A changelog is useless without directions
- Every feature worked and nobody could reach it
- Every test passed and the feature was lying
- We opened OneCamp to outside AI agents, and every gate we added found a bug
- One context window for every model was quietly wrong
- Agents that can ask each other for help — without laundering permissions
- AI teammates you can interrupt: stop it, steer it, and actually watch it work
- OneCamp AI teammates reach your whole workspace: external databases, documents, voice, and governed code
- OneCamp AI teammates open pull requests: a coding agent that ships on your own infra, not someone else's cloud
- OneCamp AI teammates get senses and skills: vision, data analysis, sandboxed code, and agents that watch your work
- OneCamp AI teammates: agents that live in your channels, with governance you can audit
- Putting a budget on AI: workspace and per-user token caps for a self-hosted workspace
- Making OneCamp programmable: a public API, scoped tokens, an MCP server, and a TypeScript SDK
- OneCamp tables: Notion-style databases your AI and API can actually use
- OneCamp collaborative boards: a Miro-class whiteboard on your own infrastructure
- OneCamp AI agents: from a chatbot that answers to a coworker that does the work
- The AI that sits in your calls: how OneCamp turns meetings into notes, decisions, and memory
- AI nudges in OneCamp: a workspace that remembers what you promised
- OneCamp connectors: I gave the workspace AI access to Gmail, Google Calendar, and GitHub
- How I built a self-hosted, AI-agnostic workspace with active memory layering
- OneCamp vs. the world: a deep technical comparison of self-hosted workspace tools
- Streaming AI in Go: SSE, circuit breakers, and the nginx buffering bug that aged me
AI Agents 7 posts
- An agent can only reach what the person asking can
- My agents kept saying done when they had only started
- Most people who tried to connect an agent failed at the first step
- Every agent row said who it acted for. None said whether anyone was watching.
- I spent a week comparing my agents to everyone else's. Then I let theirs in.
- My agents could already ask a question. They just couldn't offer the answers.
- Six things were quietly off and none of them threw an error
AI Governance 2 posts
API 2 posts
API-Resilience 1 post
Accessibility 1 post
Agencies 7 posts
- Invoices you can follow to paid
- Bill your time at the right rate
- Start a project with its plan already in it
- Tell your client where the project stands, in a minute
- Your client approves the work, then you send the invoice
- Stop writing status emails. Send your client the project
- Our clients lived in four other tools. Now they get a link
Agents 9 posts
- AI governance is mostly boring plumbing
- We opened OneCamp to outside AI agents, and every gate we added found a bug
- Agents that can ask each other for help — without laundering permissions
- AI teammates you can interrupt: stop it, steer it, and actually watch it work
- OneCamp AI teammates reach your whole workspace: external databases, documents, voice, and governed code
- OneCamp AI teammates open pull requests: a coding agent that ships on your own infra, not someone else's cloud
- OneCamp AI teammates get senses and skills: vision, data analysis, sandboxed code, and agents that watch your work
- OneCamp AI teammates: agents that live in your channels, with governance you can audit
- OneCamp AI agents: from a chatbot that answers to a coworker that does the work
Airtable Alternative 2 posts
Apps 1 post
Architecture 15 posts
- Putting a budget on AI: workspace and per-user token caps for a self-hosted workspace
- OneCamp tables: Notion-style databases your AI and API can actually use
- OneCamp collaborative boards: a Miro-class whiteboard on your own infrastructure
- OneCamp AI agents: from a chatbot that answers to a coworker that does the work
- Workflows in OneCamp: Slack-class automation without the sprawl
- AI nudges in OneCamp: a workspace that remembers what you promised
- OneCamp connectors: I gave the workspace AI access to Gmail, Google Calendar, and GitHub
- I built a full Slack-style app platform for OneCamp. Here's every decision I made.
- How I built a self-hosted, AI-agnostic workspace with active memory layering
- How I built a universal import engine supporting 8 different productivity platforms
- OneCamp: GitHub sync, custom webhooks, data archiving, and a complete frontend redesign
- OneCamp vs. the world: a deep technical comparison of self-hosted workspace tools
- Why I used MQTT for team chat (not WebSockets) - and what that taught me about broker-based architectures
- Streaming AI in Go: SSE, circuit breakers, and the nginx buffering bug that aged me
- Why OneCamp uses both Postgres AND Dgraph (and whether that was a mistake)
Asana 1 post
Async Standups 1 post
Audit 4 posts
Automation 1 post
Backups 1 post
Blockchain 2 posts
Burndown 1 post
CI 1 post
Changelog 1 post
Charts 1 post
Chat 3 posts
Check-ins 1 post
CircuitBreaker 1 post
ClamAV 1 post
ClickUp 1 post
Client Portal 3 posts
Clips 1 post
CodePR 2 posts
CodeQuality 1 post
Comparison 1 post
Connectors 1 post
Cost 1 post
Cumulative Flow 1 post
Custom Fields 2 posts
Dashboards 1 post
Data-Migration 1 post
DataSources 1 post
Database 1 post
DeadCode 1 post
Debugging 2 posts
Decompression 1 post
Defaults 1 post
Delegation 1 post
Dependencies 2 posts
Deployment 1 post
Design 2 posts
Desktop App 1 post
DevOps 2 posts
Dgraph 4 posts
Distributed Systems 1 post
Docker 4 posts
DockerCompose 1 post
Docs 1 post
Documents 1 post
Draw 1 post
Durable 1 post
EU AI Act 2 posts
Elicitation 1 post
Email 1 post
Enterprise 2 posts
Ethereum 2 posts
Excalidraw 1 post
Focus Time 1 post
Formulas 1 post
Free Plan 1 post
Freelancers 1 post
Gantt 3 posts
GitHub 3 posts
Gmail 2 posts
Go 38 posts
- Every feature worked and nobody could reach it
- Every test passed and the feature was lying
- Every health check was green and nobody could download my product
- One server per customer, and the four times the obvious design was wrong
- Two-factor, SCIM, and the config file nothing was checking
- The same one-line fault, five times, and the test that missed it too
- A failed signup burned the email address forever
- We opened OneCamp to outside AI agents, and every gate we added found a bug
- One context window for every model was quietly wrong
- The code was written. Nothing called it.
- Regex is not a sanitiser, and other things I got wrong about text
- What we fixed so you never notice: a search outage, a coding agent that couldn't run, and 31 screens that lied to you
- Agents that can ask each other for help — without laundering permissions
- AI teammates you can interrupt: stop it, steer it, and actually watch it work
- OneCamp AI teammates reach your whole workspace: external databases, documents, voice, and governed code
- OneCamp AI teammates open pull requests: a coding agent that ships on your own infra, not someone else's cloud
- OneCamp AI teammates get senses and skills: vision, data analysis, sandboxed code, and agents that watch your work
- OneCamp AI teammates: agents that live in your channels, with governance you can audit
- Putting a budget on AI: workspace and per-user token caps for a self-hosted workspace
- Making OneCamp programmable: a public API, scoped tokens, an MCP server, and a TypeScript SDK
- OneCamp tables: Notion-style databases your AI and API can actually use
- OneCamp collaborative boards: a Miro-class whiteboard on your own infrastructure
- OneCamp AI agents: from a chatbot that answers to a coworker that does the work
- Workflows in OneCamp: Slack-class automation without the sprawl
- The AI that sits in your calls: how OneCamp turns meetings into notes, decisions, and memory
- AI nudges in OneCamp: a workspace that remembers what you promised
- OneCamp connectors: I gave the workspace AI access to Gmail, Google Calendar, and GitHub
- I built a full Slack-style app platform for OneCamp. Here's every decision I made.
- Hardening a self-hosted workspace: zero-trust ingestion and client resilience
- How I built a self-hosted, AI-agnostic workspace with active memory layering
- How I built a universal import engine supporting 8 different productivity platforms
- OneCamp: GitHub sync, custom webhooks, data archiving, and a complete frontend redesign
- OneCamp vs. the world: a deep technical comparison of self-hosted workspace tools
- I spent a year building OneCamp. Here's the complete architecture.
- Why I used MQTT for team chat (not WebSockets) - and what that taught me about broker-based architectures
- Streaming AI in Go: SSE, circuit breakers, and the nginx buffering bug that aged me
- Why OneCamp uses both Postgres AND Dgraph (and whether that was a mistake)
- Server Health Monitoring Using Raft Algorithm
Goals 1 post
Governance 10 posts
- An agent can only reach what the person asking can
- My agents kept saying done when they had only started
- Most people who tried to connect an agent failed at the first step
- I spent a week comparing my agents to everyone else's. Then I let theirs in.
- My agents could already ask a question. They just couldn't offer the answers.
- We opened OneCamp to outside AI agents, and every gate we added found a bug
- Agents that can ask each other for help — without laundering permissions
- OneCamp AI teammates reach your whole workspace: external databases, documents, voice, and governed code
- OneCamp AI teammates open pull requests: a coding agent that ships on your own infra, not someone else's cloud
- OneCamp AI teammates: agents that live in your channels, with governance you can audit
GraphDB 1 post
GraphRAG 1 post
Guests 1 post
Import 2 posts
Imports 1 post
IndieHacker 1 post
Interop 1 post
Invoices 1 post
Invoicing 2 posts
Jira 1 post
Jira Alternative 1 post
Kanban 5 posts
Keyboard Shortcuts 1 post
LLM 2 posts
LiveKit 1 post
MCP 4 posts
- Most people who tried to connect an agent failed at the first step
- My agents could already ask a question. They just couldn't offer the answers.
- We opened OneCamp to outside AI agents, and every gate we added found a bug
- Making OneCamp programmable: a public API, scoped tokens, an MCP server, and a TypeScript SDK
MQTT 2 posts
Managed Hosting 1 post
Marketplace 1 post
Mattermost 1 post
Meetings 1 post
Mentions 1 post
Migration 1 post
Migrations 1 post
Mobile 1 post
Moderation 1 post
Multiplayer 1 post
NextJS 21 posts
- Two-factor, SCIM, and the config file nothing was checking
- A failed signup burned the email address forever
- One context window for every model was quietly wrong
- Regex is not a sanitiser, and other things I got wrong about text
- What we fixed so you never notice: a search outage, a coding agent that couldn't run, and 31 screens that lied to you
- Agents that can ask each other for help — without laundering permissions
- AI teammates you can interrupt: stop it, steer it, and actually watch it work
- OneCamp AI teammates reach your whole workspace: external databases, documents, voice, and governed code
- OneCamp AI teammates open pull requests: a coding agent that ships on your own infra, not someone else's cloud
- OneCamp AI teammates get senses and skills: vision, data analysis, sandboxed code, and agents that watch your work
- OneCamp AI teammates: agents that live in your channels, with governance you can audit
- OneCamp tables: Notion-style databases your AI and API can actually use
- OneCamp collaborative boards: a Miro-class whiteboard on your own infrastructure
- OneCamp AI agents: from a chatbot that answers to a coworker that does the work
- OneCamp connectors: I gave the workspace AI access to Gmail, Google Calendar, and GitHub
- I built a full Slack-style app platform for OneCamp. Here's every decision I made.
- Hardening a self-hosted workspace: zero-trust ingestion and client resilience
- How I built a self-hosted, AI-agnostic workspace with active memory layering
- How I built a universal import engine supporting 8 different productivity platforms
- OneCamp: GitHub sync, custom webhooks, data archiving, and a complete frontend redesign
- I spent a year building OneCamp. Here's the complete architecture.
NodeJS 1 post
Notion Alternative 3 posts
Nudges 1 post
OAuth 2 posts
OKR 1 post
OVH 1 post
Observability 1 post
Ollama 3 posts
Onboarding 2 posts
OneCamp 88 posts
- Docs on the demo now: toggles, tables, equations, find and replace
- A workspace with nothing in the way
- OneCamp has colour now, and it's faster where you feel it
- Three bugs the redesign turned up
- Everyone has a handle now
- Your team arrives ready to write
- An agent can only reach what the person asking can
- Move your team over, and bring everyone with it
- The demo worked. Every install didn't.
- I went through OneCamp's permission checks. Over thirty were wrong.
- Link your tables, and add up what's linked
- Formulas, sorting and filters for your tables
- Say it in a clip
- See where work piles up
- Seen: read receipts in DMs and group chats
- Dependencies for plans that overlap
- Invoices you can follow to paid
- Move to OneCamp with your fields intact
- Fields of your own, and a burndown for every cycle
- How did the week go?
- Standups without the meeting
- Bill your time at the right rate
- Goals that fill in by themselves
- Plan in hours, count time off, and bring your monday.com boards
- Tasks that wait on each other, and who has too much this week
- See every project on one page, and every plan on a timeline
- Describe a project in a sentence, start with its plan
- Start a project with its plan already in it
- Work through your tasks from the keyboard
- Tell your client where the project stands, in a minute
- Your chat, your doc and your board, side by side
- See who has what, and hand it over with a drag
- Deleting a doc said done and kept the doc
- Your client approves the work, then you send the invoice
- Sign in with your fingerprint, and other things that stopped getting in the way
- Stop writing status emails. Send your client the project
- Our clients lived in four other tools. Now they get a link
- My agents kept saying done when they had only started
- Half the team was still in Slack, and everyone's email was in another tab
- I said the server would stay closed. Today all of it is open source
- Most people who tried to connect an agent failed at the first step
- Every team had its own steps. My board had six.
- Every agent row said who it acted for. None said whether anyone was watching.
- I spent a week comparing my agents to everyone else's. Then I let theirs in.
- My agents could already ask a question. They just couldn't offer the answers.
- Six things were quietly off and none of them threw an error
- A changelog is useless without directions
- AI governance is mostly boring plumbing
- Every feature worked and nobody could reach it
- Every test passed and the feature was lying
- One dead hostname took video calls down for 26 days
- Every health check was green and nobody could download my product
- The update downloaded fine, and never ran
- One server per customer, and the four times the obvious design was wrong
- Shipping two editions, and the customer install that actually works
- Two-factor, SCIM, and the config file nothing was checking
- The same one-line fault, five times, and the test that missed it too
- A failed signup burned the email address forever
- We opened OneCamp to outside AI agents, and every gate we added found a bug
- One context window for every model was quietly wrong
- The code was written. Nothing called it.
- Regex is not a sanitiser, and other things I got wrong about text
- What we fixed so you never notice: a search outage, a coding agent that couldn't run, and 31 screens that lied to you
- Agents that can ask each other for help — without laundering permissions
- AI teammates you can interrupt: stop it, steer it, and actually watch it work
- OneCamp AI teammates reach your whole workspace: external databases, documents, voice, and governed code
- OneCamp AI teammates open pull requests: a coding agent that ships on your own infra, not someone else's cloud
- OneCamp AI teammates get senses and skills: vision, data analysis, sandboxed code, and agents that watch your work
- OneCamp AI teammates: agents that live in your channels, with governance you can audit
- Putting a budget on AI: workspace and per-user token caps for a self-hosted workspace
- Making OneCamp programmable: a public API, scoped tokens, an MCP server, and a TypeScript SDK
- OneCamp tables: Notion-style databases your AI and API can actually use
- OneCamp collaborative boards: a Miro-class whiteboard on your own infrastructure
- OneCamp AI agents: from a chatbot that answers to a coworker that does the work
- Workflows in OneCamp: Slack-class automation without the sprawl
- The AI that sits in your calls: how OneCamp turns meetings into notes, decisions, and memory
- AI nudges in OneCamp: a workspace that remembers what you promised
- OneCamp connectors: I gave the workspace AI access to Gmail, Google Calendar, and GitHub
- I built a full Slack-style app platform for OneCamp. Here's every decision I made.
- Hardening a self-hosted workspace: zero-trust ingestion and client resilience
- How I built a self-hosted, AI-agnostic workspace with active memory layering
- How I built a universal import engine supporting 8 different productivity platforms
- OneCamp: GitHub sync, custom webhooks, data archiving, and a complete frontend redesign
- OneCamp vs. the world: a deep technical comparison of self-hosted workspace tools
- I spent a year building OneCamp. Here's the complete architecture.
- Why I used MQTT for team chat (not WebSockets) - and what that taught me about broker-based architectures
- Streaming AI in Go: SSE, circuit breakers, and the nginx buffering bug that aged me
- Why OneCamp uses both Postgres AND Dgraph (and whether that was a mistake)
Open Source 2 posts
OpenSearch 3 posts
OpenSource 80 posts
- Docs on the demo now: toggles, tables, equations, find and replace
- A workspace with nothing in the way
- OneCamp has colour now, and it's faster where you feel it
- Three bugs the redesign turned up
- Everyone has a handle now
- Your team arrives ready to write
- An agent can only reach what the person asking can
- Move your team over, and bring everyone with it
- The demo worked. Every install didn't.
- I went through OneCamp's permission checks. Over thirty were wrong.
- Link your tables, and add up what's linked
- Formulas, sorting and filters for your tables
- Say it in a clip
- See where work piles up
- Seen: read receipts in DMs and group chats
- Dependencies for plans that overlap
- Invoices you can follow to paid
- Move to OneCamp with your fields intact
- Fields of your own, and a burndown for every cycle
- How did the week go?
- Standups without the meeting
- Bill your time at the right rate
- Goals that fill in by themselves
- Plan in hours, count time off, and bring your monday.com boards
- Tasks that wait on each other, and who has too much this week
- See every project on one page, and every plan on a timeline
- Describe a project in a sentence, start with its plan
- Start a project with its plan already in it
- Work through your tasks from the keyboard
- Tell your client where the project stands, in a minute
- Your chat, your doc and your board, side by side
- See who has what, and hand it over with a drag
- Deleting a doc said done and kept the doc
- Your client approves the work, then you send the invoice
- Sign in with your fingerprint, and other things that stopped getting in the way
- Stop writing status emails. Send your client the project
- Our clients lived in four other tools. Now they get a link
- My agents kept saying done when they had only started
- Half the team was still in Slack, and everyone's email was in another tab
- Most people who tried to connect an agent failed at the first step
- Every team had its own steps. My board had six.
- Every agent row said who it acted for. None said whether anyone was watching.
- I spent a week comparing my agents to everyone else's. Then I let theirs in.
- My agents could already ask a question. They just couldn't offer the answers.
- Six things were quietly off and none of them threw an error
- A changelog is useless without directions
- AI governance is mostly boring plumbing
- Every feature worked and nobody could reach it
- Every test passed and the feature was lying
- One dead hostname took video calls down for 26 days
- Every health check was green and nobody could download my product
- The update downloaded fine, and never ran
- One server per customer, and the four times the obvious design was wrong
- Shipping two editions, and the customer install that actually works
- Two-factor, SCIM, and the config file nothing was checking
- The same one-line fault, five times, and the test that missed it too
- A failed signup burned the email address forever
- We opened OneCamp to outside AI agents, and every gate we added found a bug
- One context window for every model was quietly wrong
- Regex is not a sanitiser, and other things I got wrong about text
- What we fixed so you never notice: a search outage, a coding agent that couldn't run, and 31 screens that lied to you
- Agents that can ask each other for help — without laundering permissions
- AI teammates you can interrupt: stop it, steer it, and actually watch it work
- OneCamp AI teammates reach your whole workspace: external databases, documents, voice, and governed code
- OneCamp AI teammates open pull requests: a coding agent that ships on your own infra, not someone else's cloud
- OneCamp AI teammates get senses and skills: vision, data analysis, sandboxed code, and agents that watch your work
- OneCamp AI teammates: agents that live in your channels, with governance you can audit
- Putting a budget on AI: workspace and per-user token caps for a self-hosted workspace
- Making OneCamp programmable: a public API, scoped tokens, an MCP server, and a TypeScript SDK
- OneCamp tables: Notion-style databases your AI and API can actually use
- OneCamp collaborative boards: a Miro-class whiteboard on your own infrastructure
- OneCamp AI agents: from a chatbot that answers to a coworker that does the work
- Workflows in OneCamp: Slack-class automation without the sprawl
- The AI that sits in your calls: how OneCamp turns meetings into notes, decisions, and memory
- AI nudges in OneCamp: a workspace that remembers what you promised
- OneCamp connectors: I gave the workspace AI access to Gmail, Google Calendar, and GitHub
- I built a full Slack-style app platform for OneCamp. Here's every decision I made.
- How I built a universal import engine supporting 8 different productivity platforms
- OneCamp: GitHub sync, custom webhooks, data archiving, and a complete frontend redesign
- I spent a year building OneCamp. Here's the complete architecture.
Operations 2 posts
Passkeys 1 post
Performance 2 posts
Permissions 1 post
Post-Mortem 2 posts
PostgreSQL 4 posts
Postgres 4 posts
Product 4 posts
Productivity 4 posts
Profiles 1 post
Project 4 posts
Project Management 13 posts
- See where work piles up
- Dependencies for plans that overlap
- Fields of your own, and a burndown for every cycle
- How did the week go?
- Goals that fill in by themselves
- Plan in hours, count time off, and bring your monday.com boards
- Tasks that wait on each other, and who has too much this week
- See every project on one page, and every plan on a timeline
- Describe a project in a sentence, start with its plan
- Start a project with its plan already in it
- Tell your client where the project stands, in a minute
- See who has what, and hand it over with a drag
- Every team had its own steps. My board had six.
Provisioning 1 post
Python 1 post
Read Receipts 1 post
Real-Time 1 post
Realtime 2 posts
Redis 1 post
Redux 1 post
Relations 1 post
Release 4 posts
Reliability 5 posts
- Three bugs the redesign turned up
- My agents kept saying done when they had only started
- The same one-line fault, five times, and the test that missed it too
- A failed signup burned the email address forever
- What we fixed so you never notice: a search outage, a coding agent that couldn't run, and 31 screens that lied to you
Remote Work 1 post
Reports 2 posts
Rollups 1 post
SCIM 2 posts
SDK 1 post
SQL 1 post
SSE 1 post
SaaS 2 posts
Sandbox 2 posts
Scheduling 1 post
Security 11 posts
- An agent can only reach what the person asking can
- I went through OneCamp's permission checks. Over thirty were wrong.
- Sign in with your fingerprint, and other things that stopped getting in the way
- Two-factor, SCIM, and the config file nothing was checking
- A failed signup burned the email address forever
- We opened OneCamp to outside AI agents, and every gate we added found a bug
- The code was written. Nothing called it.
- Regex is not a sanitiser, and other things I got wrong about text
- What we fixed so you never notice: a search outage, a coding agent that couldn't run, and 31 screens that lied to you
- Making OneCamp programmable: a public API, scoped tokens, an MCP server, and a TypeScript SDK
- Hardening a self-hosted workspace: zero-trust ingestion and client resilience
Self-Hosted 80 posts
- Docs on the demo now: toggles, tables, equations, find and replace
- A workspace with nothing in the way
- OneCamp has colour now, and it's faster where you feel it
- Three bugs the redesign turned up
- Everyone has a handle now
- Your team arrives ready to write
- An agent can only reach what the person asking can
- Move your team over, and bring everyone with it
- The demo worked. Every install didn't.
- I went through OneCamp's permission checks. Over thirty were wrong.
- Link your tables, and add up what's linked
- Formulas, sorting and filters for your tables
- Say it in a clip
- See where work piles up
- Seen: read receipts in DMs and group chats
- Dependencies for plans that overlap
- Invoices you can follow to paid
- Move to OneCamp with your fields intact
- Fields of your own, and a burndown for every cycle
- How did the week go?
- Standups without the meeting
- Bill your time at the right rate
- Goals that fill in by themselves
- Plan in hours, count time off, and bring your monday.com boards
- Tasks that wait on each other, and who has too much this week
- See every project on one page, and every plan on a timeline
- Describe a project in a sentence, start with its plan
- Start a project with its plan already in it
- Work through your tasks from the keyboard
- Tell your client where the project stands, in a minute
- Your chat, your doc and your board, side by side
- See who has what, and hand it over with a drag
- Deleting a doc said done and kept the doc
- Your client approves the work, then you send the invoice
- Sign in with your fingerprint, and other things that stopped getting in the way
- Stop writing status emails. Send your client the project
- Our clients lived in four other tools. Now they get a link
- My agents kept saying done when they had only started
- Half the team was still in Slack, and everyone's email was in another tab
- I said the server would stay closed. Today all of it is open source
- Most people who tried to connect an agent failed at the first step
- Every team had its own steps. My board had six.
- Every agent row said who it acted for. None said whether anyone was watching.
- I spent a week comparing my agents to everyone else's. Then I let theirs in.
- My agents could already ask a question. They just couldn't offer the answers.
- Six things were quietly off and none of them threw an error
- A changelog is useless without directions
- AI governance is mostly boring plumbing
- Every feature worked and nobody could reach it
- Every test passed and the feature was lying
- One dead hostname took video calls down for 26 days
- Every health check was green and nobody could download my product
- The update downloaded fine, and never ran
- One server per customer, and the four times the obvious design was wrong
- Shipping two editions, and the customer install that actually works
- Two-factor, SCIM, and the config file nothing was checking
- We opened OneCamp to outside AI agents, and every gate we added found a bug
- One context window for every model was quietly wrong
- The code was written. Nothing called it.
- Regex is not a sanitiser, and other things I got wrong about text
- What we fixed so you never notice: a search outage, a coding agent that couldn't run, and 31 screens that lied to you
- Agents that can ask each other for help — without laundering permissions
- AI teammates you can interrupt: stop it, steer it, and actually watch it work
- OneCamp AI teammates reach your whole workspace: external databases, documents, voice, and governed code
- OneCamp AI teammates open pull requests: a coding agent that ships on your own infra, not someone else's cloud
- OneCamp AI teammates get senses and skills: vision, data analysis, sandboxed code, and agents that watch your work
- OneCamp AI teammates: agents that live in your channels, with governance you can audit
- Putting a budget on AI: workspace and per-user token caps for a self-hosted workspace
- Making OneCamp programmable: a public API, scoped tokens, an MCP server, and a TypeScript SDK
- OneCamp tables: Notion-style databases your AI and API can actually use
- OneCamp collaborative boards: a Miro-class whiteboard on your own infrastructure
- OneCamp AI agents: from a chatbot that answers to a coworker that does the work
- Workflows in OneCamp: Slack-class automation without the sprawl
- The AI that sits in your calls: how OneCamp turns meetings into notes, decisions, and memory
- AI nudges in OneCamp: a workspace that remembers what you promised
- OneCamp connectors: I gave the workspace AI access to Gmail, Google Calendar, and GitHub
- I built a full Slack-style app platform for OneCamp. Here's every decision I made.
- OneCamp: GitHub sync, custom webhooks, data archiving, and a complete frontend redesign
- OneCamp vs. the world: a deep technical comparison of self-hosted workspace tools
- I spent a year building OneCamp. Here's the complete architecture.
SelfHosted 2 posts
Shell 1 post
Silent Failure 2 posts
Single Sign-On 1 post
Slack 2 posts
Slack Alternative 7 posts
Slash-Commands 1 post
Split View 1 post
Sprints 1 post
Startup 1 post
StaticAnalysis 1 post
TLS 1 post
TOTP 1 post
Tables 2 posts
Tasks 1 post
Teams Alternative 1 post
Templates 2 posts
Testing 4 posts
Time Tracking 4 posts
Timeline 2 posts
Tool-Calling 1 post
Traefik 1 post
Transcription 1 post
TypeScript 2 posts
UTF-8 1 post
UX 6 posts
- Docs on the demo now: toggles, tables, equations, find and replace
- A workspace with nothing in the way
- OneCamp has colour now, and it's faster where you feel it
- Every team had its own steps. My board had six.
- What we fixed so you never notice: a search outage, a coding agent that couldn't run, and 31 screens that lied to you
- AI teammates you can interrupt: stop it, steer it, and actually watch it work