Runs today · For internal teams
An AI assistant that lives in your Slack
Answers in the channel, from your own documents and repositories, with the source attached — and asks before it touches anything.
The job to be done
Teams of 5 to 50 that already live in Slack and keep re-answering the same five questions.
The answer exists — in a Notion page someone wrote last quarter, in a merged PR, in a thread from March. Finding it costs the asker ten minutes and the finder twenty. The fifth time someone asks how deploys work, the knowledge has stopped being a company asset and started being a tax on whoever is most online.
How it works
- 1
Someone asks, in the channel
A slash command or a mention opens the run with the question and the channel context. No one has to leave Slack.
- 2
It retrieves from your sources
Notion pages, repository files and connected knowledge are fetched through connectors, not guessed from the model’s memory.
- 3
It answers with the source
The reply carries a link to what it read. An answer without a citation is a guess, and it is labelled as one.
- 4
Actions wait for approval
Anything that writes — a ticket, a deploy, a message outside the channel — stops at an approval node until a human accepts it.
Works with
Each entry is a connector or channel available on the platform today.
What it costs
What the incumbent charges
Glean publishes no self-serve pricing: its pricing path redirects to a demo request and enterprise sales.
Where they are the better choice
Glean connects to 275+ applications with permission-aware indexing out of the box and enterprise controls (SSO, audit, ISO 42001). We connect to what your connectors cover, and you wire the permissions.
Source, checked 2026-10-07 — Glean site (demo-only pricing path, enterprise positioning)
On SynthHires
SynthHires
Free tier, then $9/month. You pay your model provider directly — we take no markup on tokens.
- No enterprise sales call: connect the workspace, add a key, invite the assistant.
- No per-seat multiplication — one flat platform price for the whole workspace.
- Every action is an inspectable automation definition you can edit, version and diff.
Your API keys are encrypted locally with AES-GCM and never travel to our servers.
See pricingWhere this is not the right fit
- Glean ships hundreds of pre-connected, permission-aware integrations. Ours covers the connectors in our registry, and anything else is an MCP server you bring.
- We do not index your whole company by default. You choose what the assistant can read, which is safer and also less turnkey.
What it handles
Channel-native
Slash commands and mentions route into the automation engine over Slack’s events API. The answer arrives where the question was asked.
Your knowledge, your key
Retrieval runs against your connected sources and the model key you configured. No training on your documents, no third-party index.
Scoped to what the team allows
Capability policy defines which connectors and hosts the assistant may touch. Out of scope means the tool call fails, not that it is discouraged.
Priced per workspace, not per seat
The platform price does not scale with how many people type in the channel.
Availability: this workflow is assembled from primitives that ship today. You configure the trigger, the agents and the gate; nothing is hidden behind a feature flag.
Frequently asked questions
Does it train on our internal documents?
No. Documents are read at answer time through connectors, sent to the model provider you configured, and used for that request only. The BYOK promise covers this: your keys and your content are not our training data.
What happens when it does not know the answer?
It says so and links what it did read. Retrieval with citations is the whole design: an assistant that invents an internal policy is worse than no assistant, because people act on it.
Can it post or act outside the channel?
Yes, through the connectors, and every write goes through an approval node first. The default posture is that the assistant answers in Slack and asks before it changes anything else.
Do we need an AI budget to run it?
You need a model key and the provider account behind it. The platform charges nothing for the tokens: you pay your provider at their rates, and budget caps on the run keep a bad loop from becoming a bad invoice.
Claims on this page last checked against the source on 2026-10-07.
- SynthHires pricing (free tier, $9/month) (2026-10-07)
Deploy this workflow with your own keys
Free tier, no card. Bring your provider key — it is encrypted in your browser and never reaches our servers.