fal.ai, without opening fal.ai.
One key, every model worth running, and the price before you approve.
- Free to start
- No card
- Nothing irreversible without your yes
- One key for the workspace
Ask in your own words.
Nothing to configure and no builder to open. Copy one of these into a thread, swap in your own fal.ai details, and Ray takes it from there.
@Ray run the MiniMax voice on this script and send me the file.
@Ray what would it cost to run that model on fifty images?
@Ray is that render finished yet?
@Ray cancel the job I started by mistake.
What Ray can and cannot do in fal.ai
Reading is free, because reading changes nothing. Everything that writes waits for your yes. The rest is not on Ray’s list.
5 reads wired, and 5 actions that wait for your yes.
Runs on its own
What Ray reads in fal.ai
- Which models are available, searched by name or category
- What an endpoint charges, and what a planned run would cost
- Whether a queued job has finished, and its result when it has
Reading changes nothing, so Ray never stops to ask whether it may look something up.
Asks first, every time
Always waits for your yes
- Run a model, wait on it, or queue it to collect later
- Upload an image, audio, or video file for a model to work from
- Cancel a job that has not started yet
Ray posts a card showing exactly what it is about to do. Nothing happens until someone approves, and no setting turns that off.
Not on the list
Things Ray is never handed
These are left out during curation, so they are not something Ray could do and decides against. They are not in front of it at all.
- Stream a job's status, which a plain status check answers better
- Fetch the webhook signing keys
Exactly what Ray will ask you to approve.
Not a category of permission. These are the fal.ai actions that stop in the thread with the real content before they happen.
- Cancel a fal.ai job
- Run a fal.ai model, which spends credits
- Queue a fal.ai job, which spends credits
- Run a fal.ai job and wait for it, which spends credits
- Upload a file to fal.ai, which returns a public link anyone can open
Cancel a fal.ai job
Showing Request id, each carrying the exact content that will land.
How Ray works with fal.ai
No workflow to build, and nothing to maintain when the work changes.
- 01
Connect fal.ai
fal.ai has no per-person login, so it takes one workspace key. Ray posts a connect card, you paste the key on Ray's own page rather than in Slack, and it covers the whole team from then on.
- 02
Ask in plain words
In the conversation you were already having. Nothing to build, no trigger to map, no tab to open.
- 03
Approve, and it lands
Reading runs on its own. Everyday changes land at once with an Undo, and anything that sends, spends or can't be taken back waits for your yes with the exact content in front of you.
Megan
@Ray the checkout bug is back. Take a look in fal.ai, and tell Dana it is being looked at.
Cancel a fal.ai job
- Waiting on
- your yes
Ray uses fal.ai together with the rest of your tools.
One ask can read fal.ai, cross-reference what is in these, and land the finished result back in the thread. Not four connectors taking turns, one teammate with one memory.
What does connecting fal.ai to Ray actually do?
Ray is an AI teammate that works in your team chat, on Slack today. Connect fal.ai and it can do the fal.ai part of a request without anyone leaving the conversation: read what it needs and make the change, asking first before anything that can’t be taken back.
It is not a workflow builder. There is no trigger to configure and no automation to repair when your process shifts. You describe the outcome in the thread and Ray picks the actions, including actions in your other connected tools when the request spans more than one.
fal.ai has no per-person login, so it connects once with a single workspace key and covers the whole team. Ray reaches whatever that key reaches, which is worth scoping the way you want everyone to see fal.ai. Anything irreversible still waits for a yes.
Security and permissions
- You paste the fal.ai key on Ray’s own page, never in Slack. It goes straight to the connection and never reaches the model.
- What waits for a yes is decided in the execution path, not by a prompt. No setting lets an irreversible action through without one.
- Each workspace is isolated at the database level, and that isolation is tested by trying to break it.
- An admin, or whoever connected it, can disconnect fal.ai in a message. Uninstall Ray and every connection it held is revoked.
fal.ai and Ray, answered
The questions teams ask before they connect fal.ai. Every answer is how this integration is built rather than a promise about it.
More design & whiteboards integrations
Ray uses them together in a single run when a request spans more than one.
Put Ray to work in fal.ai.
Free to start, no credit card, and nothing irreversible happens without your yes.
By adding Ray, you accept the Terms and Privacy Policy.